The system that Intel and Fujitsu implemented addressed both problems in a two-part approach. In the first part, video cameras watched device screens during final testing, while the gateway aggregated the video streams and sent them to the cloud for processing using text recognition technology to detect and recognize any error codes displayed on screen. When the system detected patterns in the reported errors, that analysis and the relevant video moved to a human operator for study. This operator then worked to determine the root cause of the problem by analyzing the circumstances that lead to the fault.
For the second part, the factory attached beacons to any devices sent to the rework department, enabling the real-time tracking of device movement by every worker in the department. The system annotated the location of each device with information on its shipping deadline, allowing workers to independently prioritize their work and to help out in processes that were causing delays.
Fujitsu recently announced the results of this trial. The analysis of error reports allowed the company to identify the misdetection of faults, which helped reduce the incidence of rework. The ability to prioritize rework in real-time resulted in a 30% reduction shipping costs by minimizing missed deadlines.
This kind of trial helps take the hype out of the IoT and grounds the benefits in hard numbers. With solid examples as a guideline, potential industrial IoT users gain the ability to perform cost/benefit analysis with confidence in the results. As the IoT proves its benefit, even the risk-adverse will start to adopt the technology, fueling growth in the industrial IoT market.