PROJECT TITLE :

Creating Realistic Virtual Textures from Contact Acceleration Data

ABSTRACT :

Trendy haptic interfaces are adept at conveying the large-scale shape of virtual objects, however they often provide unrealistic or no feedback when it involves the microscopic details of surface texture. Direct texture-rendering challenges the state of the art in haptics as a result of it needs a finely detailed model of the surface's properties, real-time dynamic simulation of complicated interactions, and high-bandwidth haptic output to enable the user to feel the resulting contacts. This paper presents a new, absolutely realized solution for making realistic virtual textures. Our system employs a sensorized handheld tool to capture the texture of a given texture, recording 3-dimensional tool acceleration, tool position, and contact force over time. We have a tendency to cut back the three-dimensional acceleration signals to a perceptually equivalent one-dimensional signal, and then we have a tendency to use linear predictive coding to distill this raw haptic info into a database of frequency-domain texture models. Finally, we render these texture models in real time on a Wacom pill employing a stylus augmented with tiny voice coil actuators. The ensuing virtual textures give a compelling simulation of contact with the real surfaces, that we have a tendency to verify through an individual's subject study.