USE OF GENETIC ALGORITHM TO OPTIMIZE SUBMARINE LQR CONTROLLER

Abstract

Optimal Control theory provides effective solution for applications in every field. Linear Quadratic Regular (LQR) is an important player in the field of Linear Optimal Control, but depends on the effective selection of Q and R Matrices. Optimization based on heuristic search technique like Genetic Algorithm can be used for the best Q and R matrices of LQR. This work presents the parameter optimization with genetic algorithm (GA) to illustrate effectiveness of GA in designing optimized LQR controllers. Depth control of a submarine is used as the case study in this study. As compared to simple LQR, the GA based LQR resulted in effective and optimized controller performance of states (downward velocity, pitch rate, pitch, and depth) and control actions (bow and stern control surfaces)