Object identification using support vector regression for haptic object reconstruction

Abstract

The lack of realistic haptic feedback has become a significant barrier to achieve realization in virtual reality. If an object is to be reproduced in the haptic dimension, it's essential to analyze the object behavior for mechanical inputs. Nevertheless, prior studies have considered model-based approaches to model the behavior of the real object for reconstruction, and the conventional spring-damper model was the most widely used. However, proper object identification is crucial in accurate haptic object modeling for reconstruction. Thus, this paper proposes an AI-based approach using a nonlinear regression algorithm, Support Vector Regression (SVR). AI algorithm predicts the object’s response for motion parameters by analyzing the nonlinear responses from the object extracted through a sensorless sensing system based on disturbance observer (DOB) and reaction force observer (RFOB). Furthermore, the viability of the proposed approach is demonstrated by comparing it to the conventional model-based approach.