Robust Encounter Avoidance and Conflict Resolution for Advanced Air Mobility (REACT-AAM)

In Phase I, we demonstrated initial feasibility of innovative improvements of our Robust Encounter Avoidance and Conflict Resolution (REACT) Technology to handle multiple simultaneous encounters expected to arise in the dense Advanced Air Mobility airspaces. Specifically, the innovations include an efficient implementation of evolutionary algorithms to produce a multi-segment maneuver for resolving encounters that would have resulted to near-mid-air collisions. nbsp;Another innovation is simultaneous risk assessment in the face of multiple encounters within transit corridors. nbsp;Phase I showed that REACT exceeds the target level of safety of 10e-8 loss of vehicle incidents per flight hour. SSCI proposes to further extend our existing REACT solution by incorporating additional realistic AAM constraints and data from an airspace information manager. nbsp;The data will be used for maneuver generation and construction of the fitness function. nbsp;To prevent cases when no feasible solution can be found, we will augment our baseline strategy by a hierarchical constraint weighting function to guarantee a feasible result. With broader context of aircraft intent from the airspace information manager, the FORECAST module will produce improved predictions of vehicle traffic patterns and potential loss of separation. nbsp;The focus will be toward the Regional Air Mobility (RAM) use case, where untowered airports and shared manned/unmanned airspace will be typical. nbsp;We plan to use the Terminal Area Encounter Model as a basis for developing realistic encounter scenarios. nbsp;Since high-risk phases of flight are in the terminal areas, our analysis during Phase II will focus on applications of our innovation to encounters typical during departure and arrival scenarios. nbsp; We will prepare a hardware-in-the-loop (HWIL) testbed and scenario generator to produce performance analysis of the EA-based avoidance guidance. nbsp;We will compare the EA-based performance to a baseline that uses single-segment maneuvers.nbsp; nbsp;