Autonomous planning of optimal four-dimensional trajectory for real-time en-route airspace operation with solution space visualisation

Abstract

One of the challenges in applying autonomous operation-based conflict detection and resolution (CD&R) techniques to practice is their functional completeness, such as the ability to simultaneously consider three-dimensional (3D) scenarios, controlled time of arrival, restricted areas, trajectory recovery, and multi-stakeholder performance preferences. Besides, the effective interaction between controllers, pilots and autonomous operating system can better support the CD&R to work in practice, where humans are able to intuitively understand the rationale for solutions given by automatic decision-making tools. To bridge the gap, a novel operational framework for four-dimensional (4D) trajectory conflict management in the context of free route airspace is proposed. Airspace discretisation based on a 3D grid is used to facilitate the consideration of restricted areas and to balance calculation speed and optimal effect. The core of this framework is a two-stage real-time autonomous 4D trajectory planning method in a restricted en-route sector with multiple flight levels enhanced by solution space visualisation. Stage one is a trajectory pre-planning model based on a visibility graph and the Dijkstra algorithm considering the conflict between aircraft and emerging restricted areas. Stage two is a real-time conflict-free and fuel-optimal 4D trajectory re-planning model. The space–time prism is employed to enable the planned trajectory to meet the controlled time of arrival and trajectory recovery requirements, where the performance preferences of both the air traffic management and airspace users are taken into account. The reachable solution space representation is applied to realise the solution space visualisation that has the potential to support the human–machine interaction and air–ground coordination. A simulation scenario is established to validate the effectiveness, efficiency, stability, and timeliness of the proposed method. Results show that in an artificial en-route sector, the average extra flight distance, average flight delay, average extra fuel consumption, and average computing time for conflict resolution were less than 0.41 km, 2.1 s, 1.5 kg, and 0.08 s in the saturated scenario, respectively, and the domino effect of the flight conflict did not show a significant rise with the increase of traffic density, which verified that the proposed autonomous trajectory operation methodology is a promising approach to facilitate autonomous air traffic management in the future.