Affiliations: Wuhan University of Technology, Wuhan, Hubei, China | Lehigh University, Bethlehem, PA, USA
Note: [] Corresponding author: Xiaohua Cao, Wuhan University of Technology, Wuhan, Hubei, China. E-mail: [email protected]
Abstract: In a mixed-product assembly line, material supply is very complicated, dynamically changing and error-prone due to large material flow. These uncertainties often cause some imperceptible Abnormal Logistics States (ALSs), which seriously hinder accurate and efficient delivery of materials to the assembly lines. There is no relevant theoretical model and efficient information collection technology so far. This paper aims to detect those imperceptible abnormal logistics states in mixed-product assembly lines through a novel RFID-based multi-deviation detection approach. Some ALSs' parameters are defined from different perspectives such as time, location, number, sequence and path. A calculation model of ALSs has been built by processing RFID data, then a RFID-based multi-deviation model is presented to quantify the abnormal degree of ALSs. Based on the developed models, a judging algorithm for ALSs of material supply is proposed. Compared to conventional eKanban monitor system, the proposed approach can detect the ALSs of material supply efficiently and accurately, including those that are difficult to detect in eKanban monitor system.
Keywords: Anomaly detection, abnormal logistics states, RFID, material supply
