Assessment of the impact of video aid in assembly operations measured with learning curves
The aim of the project is to formulate a system of experiments that can control a wider set of variables than previous systems. Including different training modalities.
Background
Videos seem an efficient media to communicate procedural instructions from expert to untrained operators. There are many scientific articles claiming that providing videos will help operators in achieving their procedural tasks, and further studies have begun to propose augmented reality instructions as even better than videos; however, there is a lack of studies to assess the impact of video aid on assembly tasks based on solid statistical data. My research aims to solve this gap and investigate the effectiveness of video aid in enhancing the ability of trained and untrained industrial operators to perform novel assembly tasks. In order to achieve the goal, it proposes the use of a standard model in manufacturing, the learning curve, and adopts state-of-the-art statistical methods to compare learning curves under different experimental conditions, i.e. when video aid is given or not to the operators.
This is important because training operators implies training costs. A sustainable development should reuse resources, including data and knowledge, when possible. Videos are considered a good media to store procedural knowledge, but their effectiveness in knowledge transfer stands unverified. The transition to an industrial internet of things relies on the successful ability to create digital know-how and added value for the industry.
Aims and objectives
Given the conceptual framework, this activity has found out that there is potential in the use of functional statistical analysis of variance for the comparison of learning curves. Furthermore, the experimental results provide insights on the usefulness of videos for the transfer of short-term novel instructions, but not for the long-term learning. This suggests that different training modalities have to be implemented under different conditions: when there is a need for a fast switch of tasks, a video is effective in conveying new knowledge that can be immediately exploited; on the other hand, when the operations can be planned ahead, a training that relies on independent learning and the interaction with experts is far more effective.
Project plan
Formulate a systematic of experiment to control a wider set of variables compared to the previous ones. Including different traing modalities.
Applied interdisciplinarity
The main discipline concurring to this effort are three: cognitive theory associated to the process of learning, human-computer interaction through visualization and assembly technology. For the second phase the involvement of ITM and KTH colleagues with specific knowledge in the first two domains is fundamental. The somehow surprising result of the pre-study needs to be discussed and analysed jointly to design a meaningful second phase.
Papers
- Measuring the effect of automatically authored video aid on assembly time for procedural knowledge transfer among operators in adaptive assembly stations. Andrea de Giorgio, Malvina Roci, Antonio Maffei, Milan Jocevski, Mauro Onori, Lihui Wang. International Journal of Production Research, 2021.
KTH Collaborations
The IIP department units of Digital Smart Production (leader prof. Antonio Maffei) and Sustainable Production Systems (leader prof. Lihui Wang) at KTH
The KTH course Tillverkningsteknik (MG1026) led by Mat Bejhem and the KTH course Assembly Technology (MG2040) led by prof. Antonio Maffei
INDEK, KTH
Other collaborations
Mechanical Department of the Polytechnic University of Milan (statistic analysis)
Duration
January 2019 - ?
Project participants
