Identification of haptic exploration procedures over textile surfaces with a Leap Motion Controller
DS 91: Proceedings of NordDesign 2018, Linköping, Sweden, 14th - 17th August 2018
Year: 2018
Editor: Ekströmer, Philip; Schütte, Simon and Ölvander, Johan
Author: Gussen, Lars C.; Ellerich, Max; Schmitt, Robert H.
Series: NordDESIGN
Institution: Werkzeugmaschinenlabor der RWTH Aachen
ISBN: 978-91-7685-185-2
Abstract
The development of high-quality products, which simultaneously address the customer’s needs, is a key challenge for companies nowadays. Besides features and technology a customer assesses a product by its sensory characteristics, i.e. primarily on the basis of visual, acoustics and haptic perception. Although the general importance of the sensory design of products has been recognized by industry, methodical aspects regarding the realization thereof are still insufficient in certain areas. Besides e.g. the simulation of sensory perception, the reproduction of customers’ habitus to sensorily approach and explore products is still not fully understood. In this context, especially the manner of haptic exploration of product surfaces plays an important role during the overall assessment of quality. Knowing the haptic exploration is necessary for the technical replication of customers’ perception e.g. with special haptic sensors. Haptic characteristics of products or materials are explored performing specific movements of the hands, so-called exploration procedures. An exploration procedure is a movement pattern, which is motivated by the object properties such as shape, size and surface. The aim of the work is to devise an automatic system, which is able to record specific surface exploratory procedures and to effectively identify representative gestures by means of machine learning. It is assumed that customers explore surfaces in their own way; however, the gestures they use are similar between customers to a great extent and can therefore be clustered into homogeneous groups of gestures. For recording the exploratory movements and the human surface interaction, a Leap Motion Controller by the American company Leap Motion, Inc. is applied. To investigate the usability of the Leap Motion Controller for the intended aim, two empirical studies are conducted by asking subjects to explore a textile surface. The extracted data from the controller of the first study is used to define groups of gestures. The second study is used to train different algorithms to assign the executed exploration movements to the predefined groups of gestures. The results show that the developed method is effective, i.e. it is possible by means of machine learning to show that customers use the same exploration gestures for material surfaces. Knowing how customers approach a material surface enables e.g. product design departments to reproduce customers’ habits and to address the customers' haptic perception to their interests
Keywords: Sensory Design, Human Material Interaction, Haptic Perception, Exploration Procedure, Hand Gesture Recognition, Leap Motion Controller, Machine Learning, Textiles