Comparison of Back Stepping Optimized via PSO Algorithm and LQR Controllers for a Quadrotor

Comparison of Back stepping method optimized via particle swarm optimization algorithm and LQR method for hovering control of a quadrotor is presented in this paper. Quadrotor is not a stable dynamical system and development of high performance controllers for it is important. First the dynamic model of a quadrotor is introduced and state-space equations are presented in order to simulate the dynamic model. Then two Back stepping and LQR controllers are designed to control Euler angles and height of the quadrotor. In order to optimize back stepping controller, its parameters are determined using particle swarm optimization algorithm to minimize cost function considered for LQR controller. Also commands to the motors are calculated and plotted to show the feasibility of the controller. To obtain better comparison, the cost function is calculated for different weighting matrices of Q and R for two controllers and the results are compared. The results show that Back stepping controller has more ability to minimize the cost function in comparison to LQR and the cost function in Back stepping has less values for several choices of weighting matrices.