Image-based Visual Servoing Control of a Quadcopter Air Vehicle

Abstract

ABSTRACT This article proposes two optimal controllers that stabilize the altitude, attitude, heading, and position of the quadcopter in space. The first optimal control method consists of computing the proportional controller gain by minimizing the error between measured and the desired feature vector values, as opposed to the classical visual servoing (VS) scheme that requires a known value of the proportional controller gain which is often chosen empirically. The second control method relies on the optimal tuning of PD controller gains applied on the altitude, attitude, and position loops. This, however, may lead to a convergence problem, which is overcome by the use of an optimization technique, the bat algorithm (BA). This latter is an efficient metaheuristic algorithm that has been successfully used in many optimization problems. Several simulations are run in MATLAB, in which visual servo control of the Vertical Takeof and Landing (VTOL) type of Unmanned Aerial Vehicles (UAVs), known as the quadcopter, is implemented. The obtained results show that the Bat algorithm renders the classical PD controller adaptive and provides a high flexibility in tracking the UAV trajectory even in the presence of disturbances. It also guarantees the stability and accuracy of the optimal, minimal time, vision-based controller.