Continuous-curvature Dubins-inspired L 1 guidance with sliding mode control-based tracking for UAVs

This paper proposes a nonlinear guidance law for unmanned aerial vehicles (UAVs) that generates smooth and time-efficient paths, while ensuring adherence to the minimum turn radius constraint of the UAV. This strategy introduces a two-phase approach to UAV path following by adopting the classical L 1 Guidance logic to follow segments of the time-optimal Dubins path before converging to the desired path. In this way, the guidance law retains the C 2 -continuous nature of the L 1 Guidance law, while reducing the convergence time significantly and bringing it closer to the time-optimal solution, while also respecting the curvature bounds of the vehicle and utilizing the maximum allowable curvature used in Dubins path. The proposed strategy also overcomes the problem faced by constant L 1 Guidance logic in converging to loitering paths from far-off initial positions by converging to segments of the optimal Dubins path before converging to the desired path. The dynamic feasibility of the generated trajectories is finally validated by tracking the generated trajectories using a Sliding Mode Controller (SMC), applied to the 6-DOF dynamic model of a quadrotor. The simulation results demonstrated smooth curvature profiles with significantly reduced convergence time in comparison to the classical L 1 Guidance, along with a good tracking performance with quadrotors, which confirms the utility of the proposed guidance and tracking framework for UAV path following. The effectiveness of the proposed guidance logic is further validated through real-time flight experiments.