Frederick Wieland

The Design and Optimization of a Combined Arrival-Departure Scheduler

This paper investigates the concept of combined arrival-departure scheduler that assigns the time of arrivals and departures at an airport with the aim of minimizing total delays. The paper presents a Mixed Integer Linear Program (MILP)-based scheduler model that calculates the optimum sequence of aircraft in arrival and departure stream at an airport constrained by minimum wake vortex separation and their estimated time of operations. Rather than treating the arrival and departure streams as two separate entities, the paper explores the concept of solving for a better solution by modeling the two streams together by introducing additional constraints. The paper compares this approach to current practice of filling arrival and departure slots in first come first serve order using San Diego International airport as a test case. The paper presents the results from the combined arrival-departure scheduler and compares them to a first-come-first-serve schedule. Next, the paper uses the scheduler to compute the effects of a “best-equipped, best-served” concept on airport delays due to changes in arrival and departure sequences.

Looking into the Future of Air Transportation Modeling and Simulation: A Grand Challenge

The future of air transportation modeling and simulation represents a grand challenge for both the air transportation and the modeling and simulation communities. In addition to its size and scope, air transportation has many technical, operational, organizational, policy, security and legal aspects contributing to its complexity, yet must ensure an un-rivaled level of safety throughout any transition to next generation systems. Thus, modeling and simulation must address not only technical challenges in representing a system of this complexity, but also institutional realities that prevent full data and model sharing between the many entities involved in air transportation. We outline the attributes of a shared repository of data, models and computational tools required to meet this challenge, including the coordination and administrative processes required to manage it, and provide an example of its potential application.

UAS Demand Generation Using Subject Matter Expert Interviews and Socio-economic Analysis

Two approaches employed in developing traffic demand estimates for nineteen civilian and commercial applications (“missions”) of unmanned aircraft systems (UAS) are described in this paper. The missions were sourced from RTCA’s DO-320 report and selected using two criteria: 1) UAS flights are operated within continental US (CONUS), and 2) the flights present the maximum potential for interaction with commercial airline traffic in CONUS. The first approach used input from subject matter experts (SMEs) via email and phone interviews to gather information on the state-of-the-art in conducting the missions, operationg costs, constraints involved, and regulatory, operational and economic challenges pertaining to the use of UAS. Information obtained from survey of scholarly literature and other publicly available data sources was used in conjunction with SME input. The second approach involved socio-economic analysis using the Transportation System Analysis Model (TSAM), and relied on socioeconomic data such as population census, demographic distribution and income distribution. The paper also presents a Java-based interactive computer tool to help users develop tailored traffic data from the demand estimates, using criteria such as geographic area of operation of UAS flights, cruise altitude, cruise speed and flight duration.

Preface to Special Issue on Air Transportation

Our modern economy and lifestyle not only is enabled by air transportation, but also demands that air transportation maintain a large annual growth. Growing air transportation, however, will not be easy. Some countries are facing gridlock and require substantial improvements in capacity, while others need to implement their first-ever comprehensive air traffic management system and ensure that it is compatible with world-wide travel. In addition, military systems need to be integrated, and spaceflight operations as well as uninhabited air vehicles will need to interact seamlessly with daily operations. Air transportation systems also need to address considerable safety and security issues. Therefore, redesigning air transportation may be the largest engineering challenge of this century – not only is it global in scope, but it requires levels of safety and efficiency that demand careful design, analysis and certification. Simulation is a vital part at all stages of this process. This special issue is the second of two focusing on simulation in air transportation, the first being published in SIMULATION, Vol. 80, No. 1, in 2004. For both special issues we sought a range of papers as diverse as the applications of simulation in air transportation. These simulations may focus on one or a few humans in humanin-the-loop experiments, or may span regions, nations or economies. A simulation may analyze the operational efficiency of a new air traffic management procedure or facility as measured by aircraft throughput, or it may estimate economic demand for air travel, or it may estimate the safety of operations currently in place or only imagined. The method of simulation might use classic formulations such as discrete-events, or new models of system dynamics such dynamic event trees and dynamically-colored petri nets, or new paradigms for viewing system behaviors such as the evolutionary and emergent representations inherent to agent-based simulation. Each has its part in the analysis and design of air transportation.

METROSIM: A Metroplex-Wide Route Planning and Airport Scheduling Tool

The research community has been actively engaged in developing integrated arrivaldeparture and surface (IADS) scheduling systems at airports for the past several years. Metrosim is targeted at the solution to a full IADS problem at a Metroplex, including all surface as well as Metroplex airspace constraints and dependencies within the Metroplex. The goal of this research is to optimize the sequencing, runway assignment, and route allocation to maximize throughput, increase safety, and minimize the environmental footprint Additionally, to be maximally useful to the Air Traffic Management (ATM) community, such systems need to compute their solution in real time (in a few minutes) to provide decision support for controllers. A practical solution to such a large and complex problem has been hitherto elusive, but this paper presents a promising approach with an initial feasibility study using data from the New York Metroplex (N90). The results show that the system outlined herein could potentially be a practical solution to this important problem.

Implementing a Combined Arrival, Departure, and Surface Scheduler for a Metroplex

One of the largest unresolved problems for increasing capacity of the National Airspace System (NAS) is to efficiently use the terminal area airspace, particularly in metroplex regions that contain multiple large airports. The theoretical maximum throughput can only be computed by considering all of the arrival, departure, and surface operations not just at one airport, but at all the airports in a metroplex. When computed, such a theoretical maximum is reduced by stochastic real-world events: variations in winds, the presence of convective weather, unplanned actions of pilots and controllers, and many other realities. The problem is further exacerbated by the fact that, to be useful, such an algorithm must necessarily provide decision support to controllers in real time. Any one of these problems is difficult enough, but their combination makes the problem seem practically intractable. In this paper we present our approach to handling these issues. Our preliminary findings, using one of the New York airports as a test case, show up to an 8% increase in performance can be achieved relative to the current system.

Future architectures for autonomous national airspace system control: Concept of operation and evaluation

We present a straw man architecture that could serve as the basis of a future National Airspace System where autonomous aircraft are controlled by autonomous controllers. Such an architecture is characterized by geofenced areas that have equipage constraints as well as aircraft volume constraints, where these geofenced areas are computed dynamically. Airspace structure is largely avoided, but instead is dynamically computed as the airspace loading changes. The dynamically computed structure is based on the airspace systems current state as well as its broadcast intention from all participants, with allowances for mishaps and other system failures. The safety margin of the dynamic structure is continuously computed to provide guidance to the control algorithms that regulate airspace loading. Details of this architecture are presented as well as its evaluation method.

An analysis programming language applied to the double delay problem in aviation

Seamless integration and interoperation of simulations has been a goal of simulationists for many decades. The goal of this research is to advance the field of simulation interoperability and our approach is for this to happen at the analysis level. In this paper, we are introducing a novel language called Predictive Query Language (PQL) that uses a federation of models to provide future predictions. It builds upon current predictive analytics systems in that it creates the data that will populate the database, rather than mining existing data. As an example of how this language works, we use the double delay problem in aviation. This example has dual purpose of showing the utility of PQL while shedding light on an important problem in aviation.

Applying a disparate network of models for complex airspace problems

Modeling and simulation in the aviation community is characterized by specialized models built to solve specific problems. Some models are statistically-based, relying on averages and distribution functions using Monte-Carlo techniques to answer policy questions. Others are physics-based, relying on differential equations describing such phenomena as the physics of flight, communication errors and frequency congestion, noise production, atmospheric wake generation, and other phenomena to provide detailed insight into study questions. Several years ago, researchers at Intelligent Automation, Incorporated (IAI) recognized that many of the physics-based aviation models, while conceptually similar, were difficult to interoperate because of varying assumptions regarding particular aspects of flight dynamics. Despite this difficulty, the aviation community routinely use these diverse physics-based models for a single coherent study. IAI researchers have since constructed an automated method for interoperating these models in a manner that produces consistent, coherent, and comparable results even with computations that otherwise use different assumptions.

Predicting Future Unmanned Aerial System Flights

This paper presents an algorithm for generating flight data sets for Unmanned Aerial Systems (UAS) from inchoate guidance provided by industry regarding UAS aircraft counts and UAS aircraft performance. The industry guidance is derived from two sources: an RTCA study published in 2010, as well as a study by the Joint Planning and Development Office in 2012. The paper describes the algorithm in some detail, including assumptions made by both the source documents and the algorithm. The paper draws conclusions about the future increase in complexity of the air traffic control system due to the presence of these projected UAS flights.