Intelligent sensing systems – Status of research at KaProm

Abstract

Within Industrie 4.0 intelligent sensing systems represent an indispensable asset with significant role in enabling shifting from automated to intelligent manufacturing. Instead of being simple transducers, intelligent sensors are able to retrieve useful information from raw signal. They represent systems with integrated computation and communication capabilities, that run sophisticated and real time applicable algorithms and communicate the necessary information to the other elements of the manufacturing facility. In this paper we present the recent research results in the field of intelligent sensing systems that were accomplished at Laboratory for Manufacturing Automation and Laboratory for Robotics and Artificial Intelligence at Department for Production Engineering (KaProm) at Faculty of Mechanical Engineering in Belgrade. Presented systems are intended for application in various manufacturing processes, such as machining, assembly, manipulation, material transport, rubber processing lines. They are based on application of different non-stationary signal processing (Discrete Wavelet Transform, Huang-Hilbert transform) and machine learning and artificial intelligence techniques (Support Vector Machines, Artificial Neural Networks, bio-inspired algorithms, clustering methods, fuzzy inference mechanisms). The most of developed systems are implemented in embedded devices and their real-world applicability is demonstrated.