The image below illustrates the functional architecture for the AACAS. The architecture is split into 4 major subsystems: Sensing, Perception, Planning, and Flight. These subsystems process data in parallel and transfer data asynchronously. Inputs for the overall system are the physical environment around the drone and mission goals determined before operation. The drone will have two operation modes: autonomous and pilot-controlled. In autonomous mode, the mission goals will be predetermined waypoints. In pilot control, the mission goals are the flight commands from the pilot. The system outputs flight advisories to the operator and drone motion.
The Sensing Subsystem, as depicted in the green box, receives and processes sensor data. Perception sensors detect physical objects near the drone; these data are sent to the Perception Subsystem. Motion and position sensors measure activity from the Flight Subsystem. These data pass through a state estimator en route to the Planning Subsystem.
The Perception Subsystem receives perception data from the Sensing Subsystem. With these data, sensor detection algorithms determine and track the position of objects relative to the drone. Finally, the system predicts the trajectory of the objects for the Planning Subsystem. In addition to locating objects, another algorithm will classify each object to influence the trajectory predictor using each classified object’s dynamic model.
Flight paths are generated in the Planning Subsystem using the predicted object trajectories and drone position. The path planner will take a desired path, determine if there is a collision risk, and divert the path if necessary. The remote operator is notified of any flight path changes. The Flight Subsystem receives and executes the flight plan.