Development of Smart Tongue using Ionic Polymer Metal Composite for Robotic Applications

Abstract

This paper focus on the development Ionic Polymer Metal Composite (IPMC) based robotic tongue as artificial taste sensor to identify and discriminate between different taste senses. In this study, taste samples are used which denotes five taste senses- Salty, Sweetness, Spicy, Bitterness and Sourness. NafionTM–117 (DuPont) substrate based IPMC due to its hydrophilic nature allow cations like Li+, Na+, K+ etc. diffuse through its perfluorinated polymer membrane penetrating Platinum (Pt)-surface electrode layer on both side of the membrane. Cations of different concentration present in natural and artificial food products. Based on the amount of diffused cations inside the polymer matrix of IPMC, various taste samples are identified. While bending force applied on IPMC strip, mobile cations and fixed anions get polarized inside IPMC polymer membrane and create potential difference in millivolts (mV). The potential difference measured across the Pt-surface electrode of IPMC. Magnitude of the developed emf depends on cationic concentration for different taste samples. Considering taste sense as identification index, bending deformation of IPMC strip has clear correlation with magnitude of emf developed across the Pt-surface electrode. K-nearest neighbor (KNN) based Classification technique is applied in developing the Machine Learning (ML) Model for prediction of different taste senses. The model is trained with real-time dataset having information about ‘output voltage developed across IPMC’ vs. ‘taste samples of variable concentration’. Training is done in two phases where the ratio of training data: test data was 70%: 30% and 80%: 20% respectively. It is observed that Prediction accuracy of the learned model increases if the volume of training dataset increases. High prediction accuracy of the developed ML model confirms the competence IPMC based e-tongue technology, applicable for robotic system as artificial taste sensor.