Proof-of-concept of a cyber-physical system for identifying container handling processes in supply chains based on accelerometers and artificial intelligence

This paper presents a proof-of-concept of a cyber-physical system for identifying material handling processes in a supply chain using artificial intelligence. Multi-axis and high-bandwidth accelerometers are used to generate raw-data to serve as input for a machine learning algorithm. The task is to classify the type of the respective movement in real-time, e.g., to decide whether the container is being manually moved by a person or by a specific means of transportation. The paper aims to provide results that will help contextualize KPIs, monitor material flows and detect transportation anomalies by gathering data gained from several sensorboxes applied in an industrial application. Thus, with the use of Internet of Things technologies and artificial intelligence it is possible to monitor and assess a logistical operation without further local infrastructure across different companies in a supply chain. The working prototype is a first step in achieving more flexibility in the tracking of container handling processes through an infrastructure-free approach that improves transparency and help with risk management in supply chains.