Profiling the behaviour of 3D selection tasks on movement time when using natural haptic pointing gestures

Proceedings of the 16th ACM Symposium on Virtual Reality Software and Technology - VRST '09 (2009), doi: 10.1145/1643928.1643947
Experiments
Tasks
Findings

How do haptic force feedback, object size, and one or two objects affect 3D object selection tasks?

Variables

    Haptic Fidelity
  • haptic quality - no force-feedback, soft force-feedback (0-3 n increasing linearly towards the center of the sphere), hard force-feedback (5 n)
Independent Variables
System Info

Displays
Input Hardware

GRAB haptic interface, a force-feedback device for the simulation of manipulation of objects with two fingers, consisting of two robotic arms and a control unit along with a visualization system.

Participant Info

All participants had similar backgrounds, had good hand-eye coordination, were right-handed, and were between 5'8" and 5'9".

Total # Age Range Gender Balance
30 20 - 23

Participants selected either one sphere or two spheres in sequence by moving their right hand to the virtual location of the sphere. Their right index finger was inside the thimble of the haptic device in order to provide haptic force-feedback when contacting a sphere.

Interaction and Environment

Interface

The user could move freely, but only their right hand, through the index finger, had haptic feedback provided, and this was all that was used to conduct the experiment. The user stood still and moved their hand to select a sphere by touching it (all spheres were within arm's reach). When a sphere was touched, it provided either no, soft, or hard force-feedback and turned gray to indicate it had been selected.

One sphere was white, and the other yellow, each placed on a gray rod. The environment depicted an outdoors scene with a fixed horizon level and a dark ground color for contrast between targets and backgound environment.

Dimensionality Scale Density Visual Realism
3D Small Low Medium
Metrics

  • time - time was measured between the start of the task and when the sphere(s) was(were) selected and turned gray.
  1. There was a significant interaction between haptic quality and number of selected objects on time for a object selection task.

    With two objects being selected, more force-feedback (hard > soft > no) reduced the time taken to select both objects. For selecting the first of two objects, no force-feedback was much faster than with one object, though there was no difference for soft

    Specificity: Somewhat specific

  2. There was a significant interaction between object size and distance of objects on time for a object selection task.

    Medium and large objects took less time to select than small ones for all distances. Large objects were selected with significantly less time only within 30 cm.

    Specificity: Highly general

  3. There was a significant interaction between number of selected objects and distance of objects on time for a object selection task.

    Selecting the first object out of two took more time than selecting the only object when only one was present, for distances less than 35 cm. Beyond 35 cm, no interaction was found.

    Specificity: Somewhat general