Haptic data is a kind of immersidata that is used to describe the movement, rotation,.
and force associated with user-directed objects in an immersive environment. We use the.
CyberGlove as a haptic user interface to an immersive environment. The CyberGlove consists.
of several sensory devices that generate data at a continuous rate. The acquired data can.
be stored, queried, and analyzed for several applications.
In this chapter, we focus our attention on the analysis of haptic data with the objective.
of modeling these data in a database. A large number of diverse applications use haptic data.
Each such application may need haptic data stored and modeled at different levels of abstraction.
For now, we consider three levels of abstraction. First, in Shahabi, Barish, Kolahdouzan,.
Yao, Zimmermann, and Zhang (2001), we made our first attempt to model haptic.
data at the lowest level of abstraction. There, we dealt with raw haptic data conceptualized.
as time-series data sets. Such a modeling approach can be used for training applications such.
as comparing a teacher's and a student's session with the CyberGlove, to measure the student's.
proficiency at following the teacher. Second, in this chapter we move a level up from.
our previous work in using raw haptic data by trying to understand the semantics of hand.
actions, and we employ several learning techniques to develop this understanding. The application.
that we focus on is limited vocabulary American Sign Language recognition that.
involves the translation of American Sign Language (ASL) to spoken words. Finally, for the.
third level of abstraction, there exists a class of applications that need to analyze preprocessed.
data, as opposed to analyzing raw haptic data. An example would be the application of.
detecting the grasping behavior of the hand. This application might need the speed of the.
hand in space at a certain instant of time. We intend to study this final level of abstraction as.