🤖 AI Summary
This work addresses the challenge of simultaneously tracking airspeed, altitude, and heading for fixed-wing unmanned aerial vehicles under severe wind disturbances, where conventional autopilots exhibit limited adaptability and end-to-end reinforcement learning risks unsafe control surface exploration. The authors propose a learning-based supervisor that preserves the original autopilot architecture by selecting residual corrections to reference commands from a discrete action set, which are then projected onto an admissible command envelope before being fed into the baseline controller—ensuring it remains the sole actuation interface. The approach integrates Hamilton–Jacobi–Bellman (HJB)-inspired principles to formulate an operator-relative advantage scoring mechanism and incorporates a safety filter inspired by control Lyapunov and barrier functions to guarantee safe exploration. Evaluated in a 12-dimensional state space, the method reduces path-tracking RMSE by 86.77% (to 44.8 m) and 49.54% compared to the baseline autopilot and tabular Q-residual methods, respectively, achieving the most significant improvement in scenarios where the original system degrades most severely.
📝 Abstract
A fixed-wing UAV must hold airspeed, altitude, and heading references under wind, gusts, and turbulence, channels coupled so that correcting one can degrade another. Classical autopilots stabilize the airframe well but adapt poorly when a hard crosswind meets an aggressive turn, while reinforcement-learning (RL) policies acting directly on the surfaces concentrate exploration risk at the actuator interface. We place a learned supervisor above an unchanged autopilot rather than inside it: it selects a residual from a finite, bounded action set on the commanded airspeed, altitude, and heading; the modified reference is projected into an admissible command envelope before reaching the autopilot, which stays the only actuator-facing controller. What is new is how the residual is chosen. HJB residual scores candidates with a semi-discrete value-iteration critic in the spirit of the Hamilton-Jacobi-Bellman (HJB) equation, ranks them by a no-op-relative Hamiltonian advantage, and filters them through a control-Lyapunov- and control-barrier-inspired finite-action shield that always keeps a no-op fallback. On a shared 12-state runtime holding the plant, autopilot, and actuator model fixed, so the comparison is at the package level, HJB residual lowers mean RMS path-tracking error to 44.809 m, against 338.617 m for the baseline autopilot and 88.809 m for a tabular-Q residual, an 86.77% reduction over the baseline and 49.54% over Q-learning. The gain concentrates where the baseline fails worst and comes with a measured rise in airspeed error, so no method dominates every metric. We present this autopilot-preserving residual command-supervision design and benchmark with its trade-offs reported intact.