Mass Estimation for an Adaptive Trajectory Predictor using Optimal Control

Abstract

Air traffic predictability is paramount in the air traffic system in order to enable concepts such as Trajectory Based Operations (TBO) and higher automation levels for selfseparation. Whereas in simulated environments 4D conflictfree trajectory optimisation has shown good potential in the improvement of air traffic efficiency, its application to real operations has been very challenging due to the current lack of information sharing between airspace users. Consequently, such operations are still very limited in scope and rarely attempted in dense traffic situations. Better predictability of other traffic future states would be an enabler for each aircraft to fly its user preferred route without decreasing safety in a self-separation context. But this is not an easy task when basic aircraft parameters such as aircraft weight, performance data or airline strategies are not available at the time of prediction. In this paper the authors propose to compensate this hindrance by continuously integrating the state of the surounding traffic to improve the ownship's knowledge of other aircraft's dynamics. Specifically, conventional position (and velocity) messages, as coming from Automatic Dependent Surveillance Broadcast (ADS-B), are integrated at the ownship. Then, an optimisation problem is formulated, using optimal control theory, that minimises the error with the known states, having the parameters of study (i.e. mass) as decision variables. A scenario with two departing trajectories is used to demonstrate the effectiveness of this parameter estimation method. In it, the take-off mass of the potential intruder is estimated onboard the ownship and its impact to conflict detection and resolution is presented, demonstrating the big improvements in predictability and safety.

Conflict-Free Trajectory Optimization with Target Tracking and Conformance Monitoring

“…Besides, the conformance monitor could have an adaptive threshold which would decrease as the two aircraft get closer. Additionally, a mass estimation algorithm such as the one described in [40] could be run at each replan to learn from the intruder's past states and produce more accurate predictions, resulting in earlier and more efficient corrective actions.…”

“…Therefore, assumptions for mass and cost index strategies could be corrected using ADS-B historical data, as the authors presented in Ref. [40]. Besides, given the flexibility with which the scenarios can be defined in our tool, it could easily be extended to a wide variety of scenarios.…”

Impact of continuous climb operations on airport capacity

“…The CD&R algorithm acts on the CCO data collected which consists of simulated trajectories. The variables described in Table 2 are those that have the greatest influence on aircraft performance (Pradines and Pablione, 2007;Thipphavong, 2008;Vilardaga and Prats, 2017). There are, however, other factors that could add greater uncertainty such as flap setting, power setting and number of engines.…”