Manufacturing industries rely on a set of predefined motions that need to be trained and executed in a form that should not generate health problems, as a result of the exposure time or the ergonomics execution. A set of predetermined motions for specific tasks is used on a daily basis in large manufacturing industries and it has been extensively studied by the industrial engineering field.
Method Study (MS) is used in industrial engineering and addresses the proper standard movement execution for a well-defined task. In the past few years, solutions regarding MS had reached a stable plateau and still relyon a high workload process. On the other hand, we are assisting, on an everyday basis, to an increasing expansion of mobile and wearable solutions that are able to record Human motion. Combining the opportunities arising from the latest wearable solutions with the industrial engineering field, this project allowed to create a cost-effective accurate motion training, applied to the MS context.
The solution created to answer the proposed challenge is able to follow Human activity on the manufacturing context using a Pandlet wearable device previously developed by Fraunhofer AICOS. This device contains integrated Inertial and Environmental Units and transmits the data using Bluetooth 4.0 interface. The employee wears the device on the wrist which will automatically record data on an unobtrusive and ubiquitous philosophy. First, a reference sample is obtained from a high qualified worker in order to retrieve an approximate ideal sequence of motions for the executed task. Secondly, the acquired data of the group of employees is compared to the reference sample. Using signal processing techniques applied to time series data it is possible to present temporal and distance measurements. Scores reflecting the proximity of the movement being performed against the reference, can also be computed, leading to a fine characterization of the movements that composes a specific task. The system output is based on a web application that is able to use the infrastructure and access real-time individual and collective information.
Author: Duarte Folgado
Type: MSc thesis
Partner: Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa