The optimization of production processes and operational procedures in Industry 4.0 requires cost-effective electronic systems for data acquisition and signal processing. In particular, integrating sensors with high data rates for condition monitoring using AI algorithms has been considered costly. Wireless solutions were often not economically viable on account of their high energy consumption. As part of the KI-Predict project, a holistic approach is being developed, which enables intelligent process monitoring with direct signal processing and feature extraction through the combination of new AI methods with specially optimized, integrated hardware.
As a contribution to the project, we are developing an energy-saving sensor interface as an application-specific integrated circuit (ASIC) with built-in microcontroller (MCU) and AI processing units. Taking this development approach greatly reduces the data volumes to be transmitted and significantly increases the efficiency of feature extraction. This makes it possible to implement extremely energy-saving wireless sensor solutions as required. In addition, companies can continue to use their existing systems infrastructure. The standard sensors currently used are replaced with sensors containing AI technology, leading to higher efficiency in the data processing chain and cost savings.