[RodrigoGomes2023] Combining Different Data Sources for Industrial Process Monitoring

The historical importance, the aesthetics, the build quality, and the rarity are characteristic features that individually or collectively define a car as a classic. Keeping the authenticity of those masterpieces, i.e., maintaining them as close as possible to when they left the factory, requires expert restoration services. For this purpose, there are guidelines defined for classic car restoration.
Monitoring the classic car restoration process so that its evidence is recorded automatically is essential to reduce the man-in-the-middle need in this process controlling.
Therefore, this dissertation aims to detect the restoration processes performed on a classic car, in a restoration workshop, implementing energy consumption sensing, combining it with location and other sensing techniques (i.e., vibration).
A state-of-the-art structured literature review was done so we could make well-informed decisions according to the system’s requirements. In addition, the research made us know more about methodologies and technologies of Intrusive Load Monitoring (ILM) systems, Location systems, and Internet of Things (IoT)/Industrial Internet of Things (IIoT) systems.
After an exhaustive comprehension of the problem, analysis requirements, and development, it was possible to reach an efficient and reliable process identification system. By combining location data, energy loads identified and vibrations detection, using, along the way, Machine Learning (ML) algorithms, we can infer some of the restoration processes each classic car goes through in the workshop (Painting, Curing, Sanding, Mineral Blasting). Also, a web application was developed to support the control of the existing
sensors in the workshop.