Joseph Rios

Solving High Fidelity, Large-Scale Traffic Flow Management Problems in Reduced Time

Traffic flow management aims to minimize total delay subject to various constraints. While there are several models to achieve this aim, those with the highest fidelity are typically the most useful but also the most computationally intensive. Often these high fidelity models can take hours or days to solve a nationwide, 2-3 hour planning horizon scenario. There is a need for a computationally feasible scheduling algorithm that achieves near-optimal performance on large, nationwide data sets if there is to be a useful tool for nationwide traffic flow management. To this end, this work describes how flights from a given traffic data set are each assigned a “congestion score.” Based upon this score, the highest scoring flights (those involved in the most congestion) are selected for use as input to a well-accepted traffic flow management model. By choosing progressively larger key subsets of flights, the tradeoff between runtime and solution quality is examined. Results are promising as the scenarios studied showed that optimizing on a subset of only 20% of all flights produces a schedule that is within 1% of the global optimum from the original “master” problem, and this reduced scenario is solved about 14 times faster.

Parallelization of the Traffic Flow Management Problem

This paper describes a parallel approach to solving the national air traffic flow management problem. High fidelity, aircraft-level approaches to this problem can determine which flights should be held on the ground or in an enroute sector. The major drawback to such aircraft-by-aircraft approaches lies in the large amount of data and resulting massive problem instances to be solved for nationwide or long planning horizon scenarios. The presented approach solves twenty individual problems, one for each Air Route Traffic Control Center in the continental United States. Individual client processes coordinate with a central server to solve their respective problems over multiple iterations. Experiments are performed using recent, historic data within a nominal scenario and a weather scenario. These experiments demonstrate the potential for solving this problem using parallel approaches in greatly reduced time. The initial results show solutions that would have taken over 24 hours to obtain optimally in a monolithic system were shown obtainable to within 3% optimality in less than one hour using the parallel architecture. In addition to the runtime and delay cost analysis, delay results from successfully running a high-fidelity, nationwide traffic flow scenario are also detailed.

Delay Optimization for Airspace Capacity Management with Runtime and Equity Considerations

En route air trac management is dicult and can benet greatly from decision support tools. This paper presents a study of the eciency and eectiv eness of two practical approaches to real-time scheduling algorithms: a simple greedy scheduler and a well-studied optimal scheduler. A subset (region) of the National Airspace System is isolated to perform optimization on a manageable portion of the airspace. The schedulers are tested on realistic data sets representing trac and conditions in the corridor between Chicago and New York area airports. In particular, the optimal scheduling of igh ts with both origin and destination in that corridor is considered, while reserving sucien t airspace for other air trac. In a majority of cases, the greedy method provides sucien t (often optimal) results, while under dicult trac and weather conditions, the optimal scheduler is worth the runtime requirements due to the inability of the greedy version to nd satisfactory solutions. Further benets of an optimal scheduler are demonstrated by incorporating the concept of equity or ‘fairness’ into the scheduling decision. Design choices in implementing equity amongst the various airlines are discussed and results demonstrating the utility of these choices are provided. Equity is easily implemented for the optimal scheduler but not the greedy, and does not require signican t additional run time. Ultimately, it is shown that an equity-aware decision support tool for delay optimization can be developed to run in real-time and can benet from incorporating more than one approach depending on the complexity of the scenario.

NASA UAS traffic management national campaign: Operations across Six UAS Test Sites

NASAs Unmanned Aircraft Systems Traffic Management research aims to develop policies, procedures, requirements, and other artifacts to inform the implementation of a future system that enables small drones to access the low altitude airspace. In this endeavor, NASA conducted a geographically diverse flight test in conjunction with the FAAs six unmanned aircraft systems Test Sites. A control center at NASA Ames Research Center autonomously managed the airspace for all participants in eight states as they flew operations (both real and simulated). The system allowed for common situational awareness across all stakeholders, kept traffic procedurally separated, offered messages to inform the participants of activity relevant to their operations. Over the 3-hour test, 102 flight operations connected to the central research platform with 17 different vehicle types and 8 distinct software client implementations while seamlessly interacting with simulated traffic.

A spatial database for reroute planning

Current storage and retrieval of air traffic management data by the FAA allows for sufficient control of aircraft flows in the National Airspace System. However, these storage methods allow neither for deep historical analysis nor for use of advanced, standards-compliant tools. Another shortcoming of these storage methods is that air traffic management data are not readily accessible by all interested parties. The contributions of this paper are a new approach to describing, storing, and serving air traffic management data, and its software implementation. This approach is compliant with relevant published standards and, therefore, is easily incorporated into current and future systems. We fully develop a schema for describing reroute advisories, implement a database based upon that schema, then implement an example tool that can use this new system to provide information in a way not possible in the current system. This schema is an extension of the Aeronautical Information Exchange Model and allows for geometric queries to databases based upon the model. The example tool uses this feature by requesting previously implemented reroutes that do not intersect a supplied region of interest. The results enable a traffic manager to base a new reroute on past decisions of all traffic managers, which is not an approach used today. In todays operations, the sharing of past decisions between traffic managers given similar situations is not a formal process. The disconnect between what various managers might implement leaves room for inefficiencies in the system and helps motivate this study. As an example of the potential benefits of the system described here, there was a savings over 1000 nmi in total flight distance over all flights affected by a particular reroute advisory.

Converging upon basic feasible solutions through Dantzig–Wolfe decomposition

We derive an important property for solving large-scale integer programs by examining the master problem in Dantzig–Wolfe decomposition. In particular, we prove that if a linear program can be divided into subproblems with affinely independent corner points, then there is a direct mapping between basic feasible solutions in the master and original problems. This has implications for integer programs where the feasible region has integer corner points, ensuring that integer solutions to the original problem will be found even through the decomposition approach. An application to air traffic flow scheduling, which motivated this result, is highlighted.

Robustness of two air traffic scheduling approaches to departure uncertainty

Linear programming methods and non-linear, evolutionary algorithm-based optimization techniques have been shown to be effective in managing large-scale air traffic flow problems. However, many of these algorithms assume perfect knowledge therefore the robustness of these algorithms in the presence of uncertainties is questionable. Since real-world application of these methods require them to be effective under uncertainty (i.e. produce few unexpected capacity violations), it is critical that they are tested in such conditions. In this paper we test the effectiveness in the presence of uncertainty of a binary programming approach and a novel, fast-learning evolutionary algorithm. Specifically we change the assumed takeoff times on which these algorithms are trained, and test the resulting solutions when takeoff delays that are consistent with historical data are incorporated. Experimental results show that without uncertainty, both sets of algorithms are able to quickly produce solutions with few to no violations. In the presence of uncertainty, the performance of the algorithms degrade with respect to the amount of delay added, but are still very good. Even when uncertainty is extremely high, the expected delay is never increased more than 30%.

Unmanned Aircraft Systems (UAS) Traffic Management (UTM) National Campaign II

The Unmanned Aircraft System (UAS) Traffic Management (UTM) effort at NASA aims to enable access to low-altitude airspace for small UAS. This goal is being pursued partly through partnerships that NASA has developed with the UAS stakeholder community, the FAA, other government agencies, and the designated FAA UAS Test Sites. By partnering with the FAA UAS Test Sites, NASA’s UTM project has performed a geographically diverse, simultaneous set of UAS operations at locations in six states. The demonstrations used an architecture that was developed by NASA in partnership with the FAA to safely coordinate such operations. These demonstrations—the second or “Technical Capability Level (TCL 2)” National Campaign of UTM testing—was performed from May 15 through June 9, 2017. Multiple UAS operations occurred during the testing at sites located in Alaska, Nevada, Texas, North Dakota, Virginia, and New York with multiple organizations serving as UAS Service Suppliers and/or UAS Operators per the specifications provided by NASA. By engaging various members of the UAS community in development and operational roles, this campaign provided initial validation of different aspects of the UTM concept including: UAS Service Supplier technologies and procedures; geofencing technologies/conformance monitoring; groundbased surveillance/sense and avoid; airborne sense and avoid; communication, navigation, surveillance; and human factors related to UTM data creation and display. Additionally, measures of performance were defined and calculated from the flight data to establish quantitative bases for comparing flight test activities and to provide potential metrics that might be routinely monitored in future operational UTM systems.

Open source, 3-D terrain visualization on a mobile device

With the increase in computing power available to mobile devices has come the ability to efficiently display geographic data, which includes 3-dimensional terrain, virtually anywhere. In the work presented here, we provide a high-level overview of software for fetching, viewing, and interacting with geographic data on a mobile device. The software is a re-implementation of NASAs World Wind code targeted for Apples iOS platform. The resulting code and application programming interface are available in an open source and extensible manner, which allows other developers to build computationally efficient and more custom applications for using 3D geographic data on the iOS platform than with other popular geographic visualization software. To demonstrate the utility of the software, we describe a tool targeted for use in general aviation that is intended to increase situational awareness while en route as well as enabling improved visualization and planning capabilities during pre-flight. To accomplish this, many of the features of the World Wind iOS implementation are exercised.

Fe(3): An Evaluation Tool for Low-Altitude Air Traffic Operations [STUB]

The concepts of unmanned aircraft system traffic management (UTM) and urban air mobility (UAM) introduce high-density operations in low-altitude airspace and will change the paradigm of the traditional air traffic system. The Flexible engine for Fast-time evaluation of Flight environments (Fe) provides the capability of statistically analyzing highdensity, high-fidelity, and low-altitude traffic system without conducting infeasible and cost-prohibitive flight tests that involve a large volume of aerial vehicles. With this simulation capability, stakeholders can study the impacts of critical factors, define requirements, policies, and protocols needed to support a safe yet efficient traffic system, assess operational risks, and optimize flight schedules. This work provides an introduction to this simulation tool including its architecture and various models involved. Its performance and applications in high density air traffic operations are also presented.