A Square-Root Fitts' Law for Modeling Heterogeneous Cursor Movements
Ron Alterovitz, Ken Goldberg and Siamak Faridani
Many human-computer interfaces require users to rapidly direct a cursor to targeted areas (e.g., menu item, buttons) on a computer screen. To model cursor motions on computers using a variety of pointing devices, we present a Square-Root Law derived using kinematics and control theory. The Square-Root Law is a variant of Fitts’ Logarathmic Law, a timing model for humans performing repetitive pointing movements. Fitts measured the average time required to tap a stylus between a pair of targets in a series of repetitive trials and proposed a two parameter logarithmic law. We consider heterogeneous cursor motions where targets vary in distance and width for each trial. We derive an alternative model using kinematics and control theory first principles that results in a two parameter, square-root function. To assess the performance of the Square-Root version, we created an online Java applet and gathered timing data from over 1000 data points from web-based participants. For heterogeneous cursor motions, the results show that the Square-Root Law is a statistically significant improvement over the Logarithmic Law and the Power Law. A better understanding of human motion models in heterogeneous contexts can facilitate the design of more efficient human computer interfaces.
Figure 1: Targets for homogenous and heterogeneous cursor motion experiments