Adaptive flight control for an autonomous unmanned helicopter

For autonomous helicopter flight, it is common to separate the flight control problem into an innerloop that controls attitude and an outerloop that controls the trajectory of the helicopter. The outerloop generates attitude commands that orient the main rotor forces appropriately to generate required translational accelerations. Recent work in Neural Network based adaptive flight control may be applied to control a helicopter where the reference commands include position, velocity, attitude and angular rate. The outerloop is used to correct the commanded attitude in order to follow position and velocity commands. This however generally requires a model of the translational dynamics which has some model error. This paper introduces adaptation in the outerloop using Pseudo Control Hedging in a way that prevents adaptation to the innerloop dynamics. Additionally, hedging is used in the innerloop to avoid incorrect adaptation while at control limits. Such an approach along with correct placement of the combined poles of the linearized system mitigates inner/outer loop interaction problems and allows one to increase bandwidth in the outerloop, thus, improving tracking performance further.