Intelligent adaptive backstepping control for MIMO uncertain non-linear quadrotor helicopter systems

Designing a controller for multi-input-multi-output (MIMO) uncertain non-linear systems is one of the most important challenging works. In this paper, the contribution is focused on the design and analysis of an intelligent adaptive backstepping control for a MIMO quadrotor helicopter perturbed by unknown parameter uncertainties and external disturbances. The design approach is based on the backstepping technique and uses a radial basis function neural network (RBFNN) as a perturbation approximator. First, a backstepping controller optimized by the particle swarm optimization is developed for a nominal helicopter dynamic model. Then, the unknown perturbations are approximated based on the universal approximation property of the RBFNN. The parameters of the RBFNN are adjusted through online learning. To improve the control design performance further, a fuzzy compensator is introduced to eliminate the approximation error produced by the neural approximator. Asymptotical stability of the closed-loop control system is analytically proven via the Lyapunov theorem. The main advantage of the proposed methodology is that no prior knowledge of parameter uncertainties and disturbances is required. Simulations of hovering and trajectory tracking missions of a quadrotor helicopter are conducted. The results demonstrate the effectiveness and feasibility of the proposed approach.