Modeling and Neural Control of Quadrotor Helicopter

Quadrotor is a rotorcraft with four vertically oriented propellers. Two of the propellers spin in clockwise direction and the other two in the counter clockwise direction. For a Quadrotor Helicopter a stabilizing controller is always needed. In this book Artificial Neural Networks based Control Methodology to stabilize the a Quadrotor Helicopoter, has been explained. Firstly a mathematical model of Quadrotor is developed. A simplified approach is adopted using momentum theory, where the gyroscopic effect and air friction on machine’s body has been neglected, resulting in a simplified model which is useful in designing a controller to stabilize the machine in hover state. The proposed model is nonlinear since the rotor dynamics are function of square of motor inputs. In the controller designing, Direct Inverse Neural Network Control methodology is employed. For that matter 16,8,4-MLP, 16,16,4-MLP and 16,64,4-MLP are used to control the Quadrotor plant. There performance is compared using simulation results. Direct Inverse Control using 16,64,4-MLP gives the best performance amongst all the other considered.