Threat-Aware Trajectory Planning for Full-Scale Helicopters

Autonomous operation of full-scale rotorcraft in contested environments requires trajectory planning methods capable of responding to threats in real time while maintaining aircraft dynamic feasibility and flight-envelope constraints. This paper presents a threat-aware trajectory planning framework for full-scale helicopters that integrates real-time threat information with a motion primitives-based trajectory generation architecture. The framework converts an abstract waypoint mission into a smooth, dynamically feasible 3D reference trajectory while accounting foroperational constraints, including airspeed, acceleration, bank angle, climb/descent, and turning limitations. Threat information is received in real time as Cursor-on-Target messages through the Team Awareness Kit network and processed to determine the threat location and corresponding avoidance region. Upon detection of a threat affecting the nominal trajectory, the planner searches a precomputed library of polynomial motion primitives representing dynamically feasible helicopter maneuvers. Appropriate polynomial coefficients are selected based on the aircraft state, threat geometry, and operational constraints and integrated to construct a smooth avoidance maneuver. The resulting avoidance segment is seamlessly stitched into the nominal three-dimensional reference trajectory, enabling rapid replanning without solving a computationally expensive trajectory optimization problem from scratch. The modified trajectory is subsequently tracked using a nonlinear model predictive control framework operating in real time. The complete architecture is implemented within a modular Robot Operating System environment and evaluated using the National Research Council Canada Bell 412 Advanced Systems Research Aircraft platform. The proposed approach was first validated using the NRC Bell 412 digital-twin simulation environment and subsequently implemented and evaluated through full-scale helicopter flight experiments involving multiple threat messages transmitted through the TAK network. Results demonstrate successful real-time generation and execution of smooth threat-avoidance trajectories while preserving vehicle feasibility and mission continuity. The proposed framework provides a computationally efficient and practically deployable approach for incorporating dynamically evolving tactical information into autonomous trajectory planning for full-scale rotorcraft.