Effects of discrete transformations of controller outputs on human tracking performance

Abstract

4 groups of 8 Ss each performed a compensatory tracking task using an acceleration control system. Each group employed a different controller output transformation: categories 3, 5, 7, or continuous. Each S used 4 gain (G) levels. Both tracking accuracy and economy were measured. The number output categories (C) significantly affected the economy (p < .05) but not the accuracy of performance. The G effects were significant for both accuracy (p < .001) and economy (p < .001). Accuracy improved and economy decreased monotonically over the lower 3 gains so that there was a trade-off between the 2 performance measures; at the highest gain both accuracy and economy were degraded. Although inspection of the accuracy data suggests that as the number of output categories increases the optimal gain becomes higher, the G X C interaction was not significant. (PsycInfo Database Record (c) 2025 APA, all rights reserved)