ABSTRACT
There have been great needs in evaluating human performances for the new technologies or procedures that are proposed to be applied in air traffic control or management. The fast-time human performance model, Air Man-Machine Integrated Design and Analysis System (Air MIDAS), filled a gap in the standard methodologies of both the fast-time logistical studies and the human-in-the-loop studies. Standard approaches use an incremental increase in testing complexity, fidelity and cost from empirical results, to prototype, to full mission simulation and then field testing. Such a process is, of course, a valid paradigm. However, the rate of development of ATM systems, the tremendous economic pressure to implement and reap immediate benefits from technologies, and the significant complexity and cost of large scale distributed air-ground tests suggest the development of other, more cost effective methods of human factors research. Prominent among these is computational human performance modeling (Laughery, Archer and Corker, 2001). In this paradigm the human and the system elements of interest are represented as computational entities (or agents). These agents interact as the system elements would in actual field operations, and behaviors can be observed. The benefit accrued is that system and human characteristics can be quickly varied (e.g. based on an assumed technology change or procedural changes) and the impact of those changes can be identified in the full-system context. Such models help focus the expensive and complex simulation and field tests. In addition, performance at the edge of system safety can be explored in the computational human performance modeling paradigm.
