摘要：AbstractThis paper presents the autonomous landing performance of UAVs using Tracking-Model Predictive Static Programming(T-MPSP) guidance and dynamic inversion autopilot. T-MPSP guidance uses state constraint and control deviation across the landing trajectory and optimizes the control requirement in the outer loop whereas the dynamic inversion technique is implemented in the inner loop control computation. The landing phase is divided into approach, glideslope and flare. Approach is considered as a circular trajectory as well as a straight line level flying phase. Glideslope and flare are ramp with a constant flight path angle and exponential trajectories respectively. T-MPSP technique uses a suitable control guess history computed from a standard reference for its first iteration and then provides correction in the guess history in the successive iterations and updates the control law until tracking reaches closer to the desired output. The control command with T-MPSP algorithm is computed at grid points across the trajectory. A sliding window approach is implemented in T-MPSP to compute the control wherein once the control is computed at a grid point the window slides forward. The nonlinear Six-DOF aerodynamic model of AE2 UAV have been used for validation.
