Multi-Scale Path Planning for Reduced Environmental Impact of Aviation
By: Scot E. Campbell
Doctoral Committee: Dr. Michael B. Bragg, Dr. Victoria Coverstone, Dr. Cedric Langbort, Dr. Natasha Neogi and Dr. Don Wuebbles
Ph.D., University of Illinois at Urbana-Champaign, 2010
ABSTRACT
A future air traffic management system capable of rerouting aircraft trajectories in realtime
in response to transient and evolving events would result in increased aircraft
efficiency, better utilization of the airspace, and decreased environmental impact.
Mixed-integer linear programming (MILP) is used within a receding horizon framework
to form aircraft trajectories which mitigate persistent contrail formation, avoid areas of
convective weather, and seek a minimum fuel solution. Areas conducive to persistent
contrail formation and areas of convective weather occur at disparate temporal and
spatial scales, and thereby require the receding horizon controller to be adaptable to
multi-scale events. In response, a novel adaptable receding horizon controller was
developed to account for multi-scale disturbances, as well as generate trajectories using
both a penalty function approach for obstacle penetration and hard obstacle avoidance
constraints. A realistic aircraft fuel burn model based on aircraft data and engine
performance simulations is used to form the cost function in the MILP optimization.
The performance of the receding horizon algorithm is tested through simulation. A
scalability analysis of the algorithm is conducted to ensure the tractability of the path
planner. The adaptable receding horizon algorithm is shown to successfully negotiate
multi-scale environments with performance exceeding static receding horizon solutions.
The path planner is applied to realistic scenarios involving real atmospheric data. A
single flight example for persistent contrail mitigation shows that fuel burn increases
1.48% when approximately 50% of persistent contrails are avoided, but 6.19% when 100% of persistent contrails are avoided. Persistent contrail mitigating trajectories are
generated for multiple days of data, and the research shows that 58% of persistent
contrails are avoided with a 0.48% increase in fuel consumption when averaged over a
year.