Iterative Learning Control to Enhance Accuracy of Repetitive Maneuvers for Aerial Robots

In this study, in order to enhance the accuracy of tracking repetitive maneuvers in Unmanned Aerial Vehicles (UAVs), a learning-based control scheme is proposed. At the outset, the controller is designed based on the sliding mode control (SMC) technique. In addition, the offline PD-type memory-based iterative learning control (ILC) is used along with SMC. The purpose of using ILC method is to reduce the effect of system uncertainty on the controller and decrease repetitive errors by adjusting the input control signal to dynamics and thus, to increase the reliability of following the desired path. In the ILC scheme, the error of states is saved during the maneuvers which will be used in the subsequent iteration. Also, in order to increase flexibility of the new control structure, ILC-SMC, a multilayer perceptron (MLP) has been developed. This network is designed to extend the control signal, generated by ILC, to similar maneuvers. The inputs of this neural network are the initial conditions for starting the maneuver and the output of the neural network is a gain that is multiplied by the stored control signal ILC and produces a new control signal. This generated signal will be suitable for similar maneuvers. The Levenberg–Marquardt (LM) algorithm has been used to train the multilayer perceptron artificial neural network. This method was then used in loop maneuvers. In this simulation, the difference between the maneuvers was in the acceleration of the maneuver, the radius of the maneuver, and the initial speed of the maneuver. This reduced the tracking error for similar maneuvers without performing the training process for the ILC control component. The presented control scheme is applied to a quadrotor aerial vehicle for tracking desired trajectories and it is shown that the vehicle is able to follow the desired trajectory better than the conventional SMC in the presence of uncertainties.