An Optimization Model of Self-Paced Tracking

A model which describes human performance in a self-paced tracking task was developed based on the notion that human operators are intermittent-acting or sampled-data servo-mechanisms. The model had a functional form in terms of the probability of success and failure resulting from the execution of a manual control task such as drawing a line between fixed boundaries. The human operator was modelled as an optimizer, balancing costs and penalties of speeds and errors to achieve a maximum expected payoff. The performance of the model was evaluated by simulating a line drawing task on a digital computer. Model predictions obtained via simulation were compared with the data collected from human subjects performing the actual task in a laboratory setting. The predictions of the model were confirmed, suggesting that human operators can in fact be modelled as optimizers when performing a manual control task.