Keywords

National, Airspace, System, State-Based, Agent-Based, Uncertainty

Presentation Type

Event

Research Abstract

The National Airspace System (NAS) is comprehensively described by five functions derived from big-picture goals: (1) conflict detection and resolution, (2) controlling aircraft states, (3) traffic flow management, (4) controlling passenger states, and (5) controlling company resources. It can be reasonably assumed that these functions interact in some way; this interaction is currently unknown. A model of the entire NAS would be helpful in discovering these interactions, yet no such comprehensive model exists. To address this problem an agent-based state-based model was created in MATLAB. To date, only functions (1) and (2) were implemented. Running the model shows that there are interactions between the two functions. However, the model suffers from high computational complexities. To address this second problem, simplifications were made to the model and in turn “runtime” was reduced. The next phase is to finish creating and implementing all five functions, after which, the effects of “uncertainties” on NAS interactions (performance) will be demonstrated. Uncertainties relate to an agent’s ability—human or automation—to detect various states and implement appropriate control actions. Knowledge on the effects of uncertainties will better help to drive decisions of all scales—tactical, minute-to minute decisions as well as long term investment decisions.

Share

COinS
 

Interaction Model of Functions in the National Airspace System

The National Airspace System (NAS) is comprehensively described by five functions derived from big-picture goals: (1) conflict detection and resolution, (2) controlling aircraft states, (3) traffic flow management, (4) controlling passenger states, and (5) controlling company resources. It can be reasonably assumed that these functions interact in some way; this interaction is currently unknown. A model of the entire NAS would be helpful in discovering these interactions, yet no such comprehensive model exists. To address this problem an agent-based state-based model was created in MATLAB. To date, only functions (1) and (2) were implemented. Running the model shows that there are interactions between the two functions. However, the model suffers from high computational complexities. To address this second problem, simplifications were made to the model and in turn “runtime” was reduced. The next phase is to finish creating and implementing all five functions, after which, the effects of “uncertainties” on NAS interactions (performance) will be demonstrated. Uncertainties relate to an agent’s ability—human or automation—to detect various states and implement appropriate control actions. Knowledge on the effects of uncertainties will better help to drive decisions of all scales—tactical, minute-to minute decisions as well as long term investment decisions.