A Deep Learning Neural Network Control Approach for Quadrotor UAV Landing on a Moving Platform

A deep learning control approach for autonomous landing of the quadrotor unmanned aerial vehicle (UAV) is investigated. First, the error position dynamics concerning the moving target and the error attitude dynamics are described. Then, the force and torque controllers are developed, where the deep learning neural network (DNN) compositing of output-layers, input-layers and modular neural functions is utilized to counteract the system uncertainty. Stability analysis demonstrates that the closed-loop systems are uniformly ultimately bounded. Finally, the proposed strategy is validated through simulation examples.