Onboard Flight Dynamic Route Optimization

by João Francisco Pereira do Vale Tavares Portugal

MSc thesis presentation and discussion.

Date: 2021-Jan-19   Time: 14:30
Room: Zoom

Abstract: Nowadays the flight management systems (FMS) of an airplane can fly
it autonomously from takeoff to landing with little intervention from the
pilots assuming there are no anomalous events. However, some events require a
diversion and subsequent route replanning. This replan is non-trivial, subject
to a restricted criteria and takes attention away from the pilots and from
air traffic controllers, attention that could be spent monitoring other flight
systems. Route planning involves finding a new efficient route, communicate it
to ATC, receive approval and reprogram the flight computer to follow the new
path. Our work presents a module capable of finding a feasible flight path
while complying with all the existing restrictions whether they are space
restrictions such as weather events or no fly zones, or the airplane movement
restrictions, removing the need of pilot intervention in this area lightening
up their work as well as the air controller's work. We model the problem as an
extension to the knapsack model and we then utilize a modified labeling
algorithm created to solve the bicriteria 0-1 knapsack problem and efficiently
recalculate the routes while taking into account all existing restrictions.
The evaluation made was focused on both the accuracy of the route calculated
and the time it took to recalculate the route and present the final result.