Prediction of perceived quality through the development of a robot-supported multisensory measuring system

Abstract

Knowing how customers perceive the quality of a product before it is on the market – this is a dream of every company. Thus, this paper shows an approach to realize the prediction of perceived quality through the development of a robot-supported multisensory measuring system.Today’s products are highly sophisticated and almost interchangeable in terms of their technical properties and functional quality aspects. Therefore, the perceived quality is becoming increasingly important to product designers. It is no longer sufficed to develop a product that addresses the functional requirements of the customer, a product must at the same time correspond to the sensory perception. The sensory design of products plays an important role in the customer’s overall assessment of quality. Sensory design so far considers visual, acoustic and haptic aspects of a product mostly separately, however the human perception is inherently multisensory as well as multimodal by nature thus the combined evaluation is inevitable.The aim of this paper is the development of a robot-supported multisensory measurement system to predict the perceived quality of surface materials by means of machine learning.Literature provides approaches, which are mostly based on human studies, to record customer’s subjective perception in order to derive design specifications for the product development process. This paper pursues, in addition to the subjective studies, the approach to make the multisensory perception tangible and objective by alternative metrological solutions. The multisensory measuring system for surface material characterization fuses haptic, optical and acoustic sensors according to the human stimulus processing chain. Subsequently, the objective measurement data is linked with the subjective data via machine learning algorithms in order to predict the perceived quality. This approach can be utilized in the product development process to forecast which materials create a value for the customer.