Probabilistic 4D Conflict Detection and Avoidance Strategy in En-Route Air Traffic

EXTENDED ABSTRACT In this work, we present a real-time 4D conflict detection and conflict avoidance algorithm in enroute traffic. The algorithm is based on identifying potential conflicts and generating flight patterns across the flight envelope of an aircraft using probabilistic search strategy. Generating flyable trajectories in air congestion and in face of obstacles such as natural objects, adverse weather conditions and restricted zones is complicated by the uncertain and limited nature of models, rather this be weather patterns or pilot decision-making (or rule-based performance) of other aircraft’s pilots. In that respect, probabilistic models which embeds a) modeling and sensing errors and b) potential rational and irrational behavior from decision-makers (such as a pilot who has done a blunder) provides a natural framework to cover the uncertainty boundaries. Within our approach, we generate 4D trajectories across these uncertainty boundaries using a new RRT* implementation. These generated trajectories are then turned into flyable trajectories using B-spline flight path generation method using flight modes. From human-factors point of view, a key point is in representing the generated solutions as simple as necessary such that the commands can executed (rather this be autonomous or manual) by the flight crew within the flight envelope of the aircraft. In that respect our pilot oriented hybrid model structure serves as a natural basis for not only flight path planning, but also for flight path automation and crew guidance and decision-support. In this paper, we also show the initial efforts on the synthetic vision aids specifically tailored to our approach. The complete strategy is shown in Figure 1.