Poornima Balakrishna

Estimating Taxi-out times with a reinforcement learning algorithm

Flight delays have a significant impact on the nationpsilas economy. Taxi-out delays in particular constitute a significant portion of the block time of a flight. In the future, it can be expected that accurate predictions of dasiawheels-offpsila time may be used in determining whether an aircraft can meet its allocated slot time, thereby fitting into an en-route traffic flow. Without an accurate taxi-out time prediction for departures, there is no way to effectively manage fuel consumption, emissions, or cost. Dynamically changing operations at the airport makes it difficult to accurately predict taxi-out time. This paper describes a method for estimating average taxi-out times at the airport in 15 minute intervals of the day and at least 15 minutes in advance of aircraft scheduled gate push-back time. A probabilistic framework of stochastic dynamic programming with a learning-based solution strategy called Reinforcement Learning (RL) has been applied. Historic data from the Federal Aviation Administrationpsilas (FAA) Aviation System Performance Metrics (ASPM) database were used to train and test the algorithm. The algorithm was tested on John F. Kennedy International airport (JFK), one of the busiest, challenging, and difficult to predict airports in the United States that significantly influences operations across the entire National Airspace System (NAS). Due to the nature of departure operations at JFK the prediction accuracy of the algorithm for a given day was analyzed in two separate time periods (1) before 4:00 P.M and (2) after 4:00 P.M. On an average across 15 days, the predicted average taxi-out times matched the actual average taxi-out times within plusmn5 minutes for about 65 % of the time (for the period before 4:00 P.M) and 53 % of the time (for the period after 4:00 P.M). The prediction accuracy over the entire day within plusmn5 minutes range of accuracy was about 60 %. Further, application of the RL algorithm to estimate taxi-out times at airports with multi-dependent static surface surveillance data will likely improve the accuracy of prediction. The implications of these results for airline operations and network flow planning are discussed.

Airport Taxi-Out Prediction Using Approximate Dynamic Programming: Intelligence-Based Paradigm

Flight delay is one of the pressing problems that have far-reaching effects on society and the nations economy. A primary cause of flight delay in the National Airspace System is high taxi-out times (time between gate push-back and wheels-off) at major airports. Accurate prediction of taxi-out time is needed to make downstream schedule adjustments and for better departure planning, which could mitigate delays, emissions, and congestion on the ground. However, accurate prediction of taxi-out time is difficult because of uncertainties associated with the dynamically changing airport operation. A novel stochastic approximation scheme based on reinforcement learning (RL) is presented for predicting taxi-out times in the presence of weather and other departure-related uncertainties. The prediction problem is cast in a probabilistic framework of stochastic dynamic programming and solved by using approximate dynamic programming approaches (particularly RL). The strengths of the method are that it is nonparametric, unlike regression models with fixed parameters, it is highly adaptable to the dynamic airport environment since it is learning based, it is scalable, it is inexpensive since it does not need highly sophisticated surface management systems, and it can effectively handle uncertainties because of the probabilistic framework. Taxi-out prediction performance was tested on data obtained from the FAA Aviation System Performance Metrics database on Detroit International and Washington Reagan National Airports. Results show that the root-mean-square prediction error calculated 15 min before gate departure time is on average 2.9 min for about 80% of the predicted flights.

Improving quality of prediction in highly dynamic environments using approximate dynamic programming

In many applications, decision making under uncertainty often involves two steps—prediction of a certain quality parameter or indicator of the system under study and the subsequent use of the prediction in choosing actions. The prediction process is severely challenged by highly dynamic environments that particularly involve sequential decision making, such as air traffic control at airports in which congestion prediction is critical for smooth departure operations. Taxi-out time of a flight is an excellent indicator of surface congestion and is a quality parameter used in the assessment of airport delays. The regression, queueing, and moving average models have been shown to perform poorly in predicting taxi-out times because they are slow in adapting to the changing airport dynamics. This paper presents an approximate dynamic programming approach (reinforcement learning, RL) to taxi-out time prediction. The taxi-out prediction performance was tested on flight data obtained from the Federal Aviation Administrations (FAA) Aviation System Performance Metrics (ASPM) database on Detroit International (DTW), Washington Reagan National (DCA), Boston (BOS), New York John F. Kennedy (JFK) and Tampa International (TPA) airports. For example, at the Boston airport (presented in detail) the prediction accuracy by RL model was 14running-average model. In general, the RL model was 35–50% more accurate than the regression model for all of the above airports. Copyright

Analysis of Special Activity Airspace for Air Traffic Management

In today’s system, air traffic management decisions are often made in the absence of realtime special activity airspace (SAA) status information. Hence, planned routes for flights often avoid the SAA when, in fact, the SAA is inactive at the time the flight is projected to cross the SAA. This avoidance may prove expensive in terms of flight time and distance, consequently impacting fuel burn and economic objectives. Through improved situational awareness and information sharing, the Next Generation Air Transportation System is expected to provide flight planners with dynamic SAA status updates. Improved knowledge of SAA availability may allow flights to plan direct routes through an SAA. In this paper, we model the impact of special use airspace (SUA), a subset of SAA, on flight trajectories in the national airspace system, and quantify benefits of improved airspace availability on flight times and distance. Using detailed SAA activity data and flight trajectory simulations, we analyze several future air traffic scenarios to quantify these benefits. Results suggest that while SUAs may impact only a small proportion of daily flights in the NAS, the savings in total flight time and distance to these specific flights may be substantial. I. Introduction ROWING demand for air travel is expected to increase congestion and delays in the National Airspace System (NAS). The need to balance demand and capacity while maintaining or enhancing efficiency and safety is, hence, of interest to multiple stakeholders. The Next Generation Air Transportation System (NextGen), an ongoing effort to transform the NAS in the mid-term and beyond, aims to accommodate the increased demand of future years. Among a wide variety of capabilities, NextGen aims to provide improved situational awareness for both strategic and tactical air traffic management decisions. NextGen improvements influence both the user and service provider domains. An effective transition to NextGen is possible, when the impact of these improvements on NAS stakeholder objectives is understood. In this paper, we present an analysis of special activity airspace (SAA) for air traffic management (ATM). In today’s system, SAA status updates may not be readily available to flight planners. Through information sharing, NextGen accounts for real-time SAA status information, enabling efficient ATM decisions. NextGen presents a strong motivation for understanding SAA benefits in the context of flight planning. In this paper, we use detailed SAA location and activity data to analyze the impact of SAA status updates on flight plan trajectories across the NAS. The organization of the paper is as follows: In Section II, we present a brief background and a review of relevant literature. Section III describes the input data for the analysis and Section IV details the method of analysis. We discuss results and observations from the analysis in Section V, and present conclusions and future work in Section VI.

Learning-based models for estimating airport taxi-out time using approximate dynamic programming

➣ Summary: ➣ The stochastic dynamic programming approach to taxi-out time predictions ➣ Provides a method for sequential taxi-out time predictions in real-time. ➣ Predicts at least 15 minutes before scheduled pushback time of a flight. ➣ Method provides increased accuracy when compared with regression-based model. ➣ Future Work: ➣ Expansion of the state space for the taxi-out time estimation (TOTE) model, to include features such as day of the week and forecasted weather information. ➣ Use of TOTE as inputs to a decision making problem for departure sequencing and downstream adjustments. ➣ Use of TOTE to estimate taxi-out times for GDP flights with specifically allotted EDCTs (Estimated Departure Clearance Times) ➣ Sensitivity of the TOTE algorithm to the look-ahead window (currently set as 15 min)

An integrated approach for the estimation of spherical form tolerance

The functionality and interchangeability of a product are key concerns during quality inspection, necessitating research to improve accuracy of inspection while reducing time and costs. The computation of form tolerance invariably involves a discretization of the surface (form) and a subsequent comparison of deviations from an evaluated “best” fit. In this paper, the effects of and interactions between various factors involved in spherical form verification—fitting algorithms, sample sizes, and sampling strategies—are analyzed. The principal objective is to consider sampling and fitting in an integrated manner and make pilot conclusions that would serve as the basis for developing decision support for part inspection. Sample sizes of 16, 64, and 256 are chosen to include a low, medium, and large sample size, respectively. Sampling strategies investigated include a randomly generated sequence, a Hammersley sampling strategy, and the aligned systematic sampling scheme. The linear least-squares fitting algorithm and a linear and nonlinear optimization approach are considered. In addition, minimum zone sphericity is computed by taking advantage of the robust properties of support vector regression, and the method is evaluated against traditional algorithms. Subsequently, these various factors are incorporated into an experimental design model, and interactions and main effects are analyzed.