My research is in the design, analysis, implementation, and evaluation of practical algorithms for air transportation systems to help transform the system and make better decisions in the face of increasing traffic. This research is important because of the high costs of delays and pollution today, as well as the projected doubling in air traffic over the next fifteen years. The introduction of autonomous aircraft into the airspace system will also present new challenges in air traffic management. For my recent research interests, please see the DINaMo website.
A high-level description of some of my research interests appeared in the 2007–2008 issue of Aero-Astro, the annual report/magazine of the MIT Aeronautics and Astronautics Department. My research has been supported by the NSF, FAA and NASA.
My research in developing algorithms for air transportation systems spans four broad topics:
- Resource allocation
- Airport congestion control. Aircraft taxiing on the surface contribute significantly to the fuel burn and emissions at airports. Our research identifies opportunities to reduce airport congestion, designs and field-tests surface management strategies, and evaluates the impacts of these and other surface management strategies.
A non-technical summary of our work at Boston's Logan airport appeared in the 2009–2010 issue of Aero-Astro, and in the Boston Globe.
- Large-scale Air Traffic Flow Management. The long-standing challenge of air traffic flow management is to efficiently and robustly optimize the trajectories of all aircraft in the system on a given day, even in the presence of uncertain capacities. We have developed a new algorithm to solve very large-scale air traffic flow management problems in a fast and scalable manner. Given flight-specific operating and delay costs, our method determines optimal trajectories in space and time in the presence of network and flight connectivity constraints as well as uncertain airport and airspace capacities. Our algorithm solves problems that are significantly larger than the state-of-the-art (triple the number of flights with ten times as many airports), more complex (no rerouting restrictions), of higher fidelity (3x finer time-discretization), and generate routes under a large number of uncertainty realizations. We can solve the deterministic problem in 5 minutes, and the stochastic variant in 20 minutes.
- Multi-objective scheduling algorithms. Scheduling takeoffs and landings on runways is challenging because it needs to address three competing considerations: efficiency, safety, and equity among airlines. A natural approach to runway scheduling is Constrained Position Shifting (CPS), which requires that an aircraft's position in the scheduled sequence not deviate significantly from its position in the first-come-first-served sequence. In this research, we have developed a new family of scalable dynamic programming algorithms for runway scheduling under CPS and other operational constraints, for a range of objective functions.
- Mechanisms for Collaborative Decision Making. Effective resource allocation requires that the agents involved be incentivized to participate and share information truthfully. In this work, we design and analyze market-based mechanisms for slot exchanges between airlines, and evaluate the nature of incentives for airlines to participate and to report their true preferences, as well as the susceptibility of different mechanisms to manipulation.
- Robustness in the presence of uncertainties
- Robust routing of air traffic flows. Convective weather (thunderstorms) is responsible for large flight delays and disruptions. In our research, we develop machine learning algorithms to translate raw convective weather forecasts into probabilistic forecasts of whether or not a route will be blocked, and to then modify routes dynamically to optimize the expected capacity of the terminal-area.
Here is a short opinion piece on this topic that I contributed to the New York Times.
- Network control algorithms. In this research, we learn network models of airport operations from surface surveillance data, and solve the optimal network control problems efficiently using approximate dynamic programming. This approach can effectively model operational uncertainties, and also address practical resource constraints such as limited gate availability.
Along similar lines, we have also developed realistic distributed feedback control techniques that guarantee that the aircraft queues in each airspace sector, which are an indicator of controller workload, are kept small. Our approach provides the first distributed feedback control strategy for realistic, multi-airport settings. We also showed how our methods could be used to mitigate the impact of weather disruptions.
- Prediction of air traffic delays. In this research, we have developed a new air traffic delay prediction model that incorporates both temporal (time-of-day, day-of-week, etc.) and network delay states (the overall condition of the National Airspace System or NAS) as explanatory variables. We use clustering to identify "typical" delay states and types of days, and use this information in our prediction algorithms.
- Human-automation integration
- High-confidence control algorithms
A fundamental design decision in the development of the Next Generation Air Transportation System (NextGen) is the level of decentralization that balances system safety and efficiency. New surveillance technologies that use satellite navigation, such as Automatic Dependent Surveillance - Broadcast (ADS-B), can potentially be used to shift air traffic control to a more distributed architecture; however, channel variations and interference with existing secondary radar replies can affect ADS-B systems. The use of data transmitted from aircraft for surveillance also raises concerns about vulnerability to GPS jamming and spoofing.
- Hybrid communication and control protocols. In recent work, we have developed a framework for managing arrivals to an airport using a hybrid centralized/distributed algorithm for communication and control. We design and simulate a protocol that combines centralized control in congested regions with distributed control in low traffic regions, and show that its performance is comparable to fully centralized strategies, despite needing only 50% of the ground infrastructure cost.
- Secure and fault-tolerant air traffic control algorithms. We have proposed a framework for secure and fault-tolerant air traffic surveillance in the presence of adversaries. Our approach fuses onboard inertial sensor information with data received from neighboring aircraft to verify position estimates and evaluate the uncertainty of the surveillance estimates. We then design a control algorithm that minimizes flight times while meeting safety constraints in adversarial environments. Using simulations, we demonstrate that the proposed algorithms are capable of adapting system operations to be robust to malicious faults, including sophisticated GPS attacks.
Other topics
mit.edu