Fuzzy systems for modelling and control in aerial robotics

We discuss state-of-the-art modelling and control techniques in aerial robotics using fuzzy systems. Unlike traditional control techniques, whose performance is totally reliant on the availability of accurate mathematical models, fuzzy systems can work reliably in the absence of complex mathematical models. In addition, fuzzy systems can also be employed to accurately represent the dynamic behaviours of the plants. This will clearly be advantageous, especially for unmanned aerial vehicles, whose dynamics are very fast and highly non-linear, coupled with severe uncertainties (e.g., wind gusts, varying payloads, and operating points as well as unmodelled behaviours). In terms of aircraft platform, our focus is on the multicopter Unmanned Aerial Vehicles (UAVs), namely, the AR.Drone Quadcopter and the Tarrot 680 Hexacopter; mainly due to their robustness, safety, and suitability for human interactions as well as indoor applications. We compare the performance of our fuzzy control systems with respect to the accuracy of the conventional PID controllers. Our research highlights the efficacy of the proposed fuzzy systems for modelling and control in aerial robotics.