A Discrete Movement Model For Cursor Tracking Validated in the Context of a Dual-Task Experiment

摘要

Understanding human cursor tracking behavior is essential in understanding human motor control.Though tracking has been hypothesized as a sequence of discrete movements, better data is needed tosupport the theory. By analyzing moment-to-moment tracking data, this paper shows that discrete, nonballisticmovements exist throughout a tracking task, and that these short submovements can becharacterized by either Fitts' law or a linear model. A cognitive model was built to incorporate thecharacteristics of these discrete movements into a dual task. Using parameters estimated through linearregression of the movement data, the model achieves a good fit to the overall performance measures ofthe dual-task experiment. This research investigates the characteristics of human motor control intracking tasks, improves modeling techniques by providing a new method for estimating trackingparameters, and advances the science of motor control with new evidence for the discrete movementtracking hypothesis. The discrete movement model presented here offers an excellent alternative toestablished control theory models that are used to simulate steering in cognitive models of driving.