In this paper, a grounding framework is proposed that combines unsupervised and supervised grounding by extending an unsupervised grounding model with a mechanism to learn from explicit human teaching. To investigate whether explicit teaching improves the sample efficiency of the original model, both models are evaluated through an interaction experiment between a human tutor and a robot in which synonymous shape, color, and action words are grounded through geometric object characteristics, color histograms, and kinematic joint features. The results show that explicit teaching improves the sample efficiency of the unsupervised baseline model.
Nowadays, shorter and more flexible production cycles are vital to meet the increasing customised product demand. As any delays and downtimes in the production towards time-to-market means a substantial financial loss, manufacturers take an interest in getting the production system to full utilisation as quickly as possible. The concept of plug-andproduce manufacturing systems facilitates an easy integration process through embedded intelligence in the devices. However, a human still needs to validate the functionality of the system and more importantly must ensure that the required quality and performance is delivered. This is done during the ramp-up phase, where the system is assembled and tested first-time. System adaptations and a lack of standard procedures make the ramp-up process still largely dependent on the operator's experience level. A major problem that currently occurs during ramp-up, is a loss of knowledge and information due to a lack of means to capture the human's experience. Acquiring this information can be used to simplify future ramp-up cases as additional insights about change actions and their effect on the system could be revealed. Hence, this paper proposes a decisionsupport framework for plug-and-produce assembly systems that will help to reduce the ramp-up effort and ultimately shorten ramp-up time. As an illustrative example, a glueing station developed as part of the European project openMOS is considered.
Since late 2019, a novel Coronavirus disease 2019 (COVID-19) has spread globally. As a result, businesses were forced to send their workforce into remote working, wherever possible. While research in this area has seen an increase in studying and developing technologies that allow and support such remote working style, not every sector is currently prepared for such a transition. Especially the manufacturing sector has faced challenges in this regard. In this paper, the mental workload of two groups of participants is studied during a human-robot interaction task. Participants were asked to bring a robotised cell used in a dispensing task to full production by tuning system parameters. After the experiment, a self-assessment of the participants’ perceived mental workload using the NASA Task Load Index (NASA-TLX) was used. The results show that remote participants tend to have lower perceived workload compared to the local participants.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.