Cursor trajectory analysis

Abstract

To create non-disturbing tactual feedback in human-computer interaction we want to predict the user’s goal, so that the user is helped toward the target and away from non-targets. In this paper we describe an exploration of cursor movements with an amplitude of 250 pixels, in eight different directions and with three different control devices (a mechanical mouse, an optical mouse and an optical trackball) to find characteristics of the cursor path which could be used to create a prediction algorithm on direction. The focus was on the mean curvature of and the variability between the paths in each direction. It can be concluded that on average cursor paths are rather straight in all eight directions and with all three devices. The variability of the paths depends on (1) direction; (2) friction of the control device; (3) user.