Prototyping and Metrological Characterization of a Data Acquisition and Processing System Based on Edge Computing

This paper describes a data acquisition and processing architecture developed for monitoring industrial machineries. The architecture is based on the paradigm of edge computing and is developed using single-board computers Raspberry Pi 3B+ connected to more WeMos Lolin D1 Mini microcontrollers. All the components underwent to a metrological characterization, and compensation algorithms were proposed for the bias errors, in accordance with the ISO GUM framework. The measurements' repeatability and reproducibility were performed in accordance with current standards. The measurement chain was verified with vibration measurements; tests performed on a shaker highlight a type A uncertainty of 0.35 m/s2. The SINAD was larger than 60 dB, with no meaningful dependencies from the temperature. The comparison with a common industrial acquisition system showed the validity of the proposed architecture.