Abstract
Human supported picking stations are commonly used in warehouses across various business sectors. So called goods-to-man picking processes are usually guided by simple cross-checks to minimize picking errors. These cross-checks
typically utilize binary light-barriers and are hence not able to accurately identify wrong picks or even provide early warnings to operators. In this work, we present an approach which tackles this problem by exploiting data gathered through RFID localization technologies in combination with machine learning methods. We are able to predict the picker’s intended movement at an early stage which is a first step towards the interception of potential picking errors and hence leading to an improvement of the overall picking performance and accuracy.
typically utilize binary light-barriers and are hence not able to accurately identify wrong picks or even provide early warnings to operators. In this work, we present an approach which tackles this problem by exploiting data gathered through RFID localization technologies in combination with machine learning methods. We are able to predict the picker’s intended movement at an early stage which is a first step towards the interception of potential picking errors and hence leading to an improvement of the overall picking performance and accuracy.
Original language | Undefined/Unknown |
---|---|
Title of host publication | 2017 IEEE Int. Conf. RFID |
Pages | 73-74 |
Number of pages | 2 |
Publication status | Published - 1 May 2017 |
Keywords
- Localization
- Machine Learning
- Prediction
- Early Prediction
- Neural Network