Tactile Sensing Glove-Based System for Objects Classification Using Support Vector Machine

Peripheral neuropathies affect around 2% to 8% of the adult population and are usually caused by diseases such as diabetes, hanseniasis, and alcoholism. The sensitivity loss resulting from this type of injury makes it much more difficult to the patients cope with daily activities such as writing, driving, and eating. Electronic gloves are capable of reproducing the human sensorial capacities regarding the classification and the differentiation of objects. In this paper, we present an intelligent low-cost glove using pressure sensors to classify objects by using a Support Vector Machine (SVM) algorithm. Two capacitive pressure sensors were manufactured from thin sheets of copper and a layer of Ethyl Vinyl Acetate (EVA). These sensors were attached to the thumb and the index fingers of a glove of polyamide-spandex. The obtained capacitance values, represented in the form of digital levels, were analyzed by a software developed in Matlab, which is responsible for both the serial communication and the interface using a Graphical User Interface (GUI). It was possible to observe that the measurements conducted for different objects gather in clusters, which may be distinguished with an error smaller than 2% for an 11-object-training session using an average Gaussian SVM classifier. The results allowed developing an electronic system for objects classification which can help patients with neuropathies and the emerging assistive technologies.