Adaptive Observer-Based Implicit Inverse Control for Quadrotor Unmanned Aircraft Robots and Experimental Validation on the QDrone Platform

Abstract

Taking into consideration the issue of the quadrotor unmanned aircraft robots (UARs) actuated by motors with hysteresis input, this research presents an adaptive dynamic implicit inverse control technique based on neural networks to achieve the desired trajectories. The following summarizes the primary technologies: 1) the hysteresis effect in UARs has been considered and eliminated by the proposed implicit inverse algorithms, which means a searching method for acquiring the real control signals is designed resulting in selecting to avoid constructing the hysteresis direct inverse model; 2) precise tracking is accomplished by designing an adaptive dynamic surface control (DSC) technology with enhanced state observer under the constraint that only the position data is available. In the meanwhile, the $L_{\infty }$ performance can be obtained by selecting the suitable parameters; and 3) the underactuated Drone platform has been constructed as well as the control results have implemented to confirm that the successful application of the proposed implicit inverse control algorithms.