Path steering is a primitive 3D interaction task that requires the user to navigate through a path of a given length and width. In a previous paper, we have conducted controlled experiments in which users operated a pen input device to steer a cursor through a 3D path subject to fixed path properties, such as path length, width, curvature and orientation. From the experimental data we have derived a model which describes the efficiency of the task. In this paper, we focus on studying the movement velocity of 3D manipulation path steering when one or more path properties vary during the task. We have performed a repeated measures design experiment of 8 scenarios, including a scenario in which all path properties were kept constant, 3 scenarios in which the path width, curvature and orientation varied, 3 scenarios of varying two path properties, and 1 scenario of varying all properties. The analysis of our experimental data indicates that a path of varying orientation or width results in a low average steering velocity. During a continuous steering, the joint where a change in path curvature or orientation takes place also significantly decreases the velocity. In addition, path width and curvature are highly-correlated to the average velocity of a segment, i.e. the wider a segment (or the smaller the path curvature), the larger the average steering velocity on that segment. The results of this work could serve as guidelines for designing higher level interaction techniques and better user interfaces for traditional HCI tasks, e.g. 2D or 3D nested-menu navigation.
|Title of host publication||JVRC 2010: Joint Virtual Reality Conference of EGVE - EuroVR - VEC (Felbach, Germany, September 27-October 1, 2010)|
|Editors||T. Kuhlen, S. Coquillart, V. Interrante|
|Place of Publication||Aire-la-Ville|
|Publication status||Published - 2010|