Using Meta-Heuristics In The Control Of A Non-Linear Input Delay Laboratory Helicopter System

Stabilizing controllers for non-linear input delay system of a lab helicopter are designed by first using a 3D search in the controller parameter space, the bounds of which are estimated by over-conservative controller parameters given by Lyapunov Razumikhin theory. Simulated Annealing and Genetic Algorithms are then used to find near-optimal controller gains that ensure empirical stability while minimizing an objective function consisting of design goals like settling time, overshoot, etc. Both linearized and actual non-linear models for helicopter elevation control are used with third and second order stabilizing controllers. Performance of the resulting control schemes is evaluated and compared.