Implementing Adaptive Function Allocation

Abstract

Adaptive function allocation, in which the control of tasks dynamically shifts between humans and machines, has been proposed as an alternative to traditional static allocation, in which task control is assigned during system design and remains unchanged during operations. Understanding adaptive allocation and its effects on operator performance and workload is limited by sparse systematic research and an underdeveloped theoretical framework for implementation. The purpose of this research was to determine the efficacy of adaptive allocation by implementing adaptive allocation in a multiple task aviation simulation based on a taxonomy with facets of invocation philosophy and allocation strategy. Pilot performance was evaluated to determine benefits and costs for the implementation of adaptive allocation in a multitask aviation simulation with tracking, system monitoring, and target identification tasks. The results provide support for the implementation of adaptive allocation based on a hybrid model comprising elements of operator performance and mission relevant variables. Implementation of adaptive allocation was an effective countermeasure to the predictable decrease in tracking performance associated with the initial presentation of a surface target. Benefits were also identified for monitoring task performance despite the fact that the monitoring task was not modified by the implementation of automation. Secondary benefits included improved time estimates. Potential costs of adaptive allocation included performance variability in the tracking when task partitioning was the adaptive strategy. Implications of results for design, philosophy, and theory are presented.