Extended Kalman Filter and LQR controller design for quadrotor UAVs

A quadrotor is a unique class of UAVs with vertical take off and landing (VTOL) capability and has attracted significant attention due to its importance in various applications. This thesis presents the design and experimental implementation of Extended Kalman Filters (EKFs) to estimate the states of a quadrotor and a Linear Quadratic Regulator (LQR) controller with integral action to meet the desired control objectives. In case of the Extended Kalman Filters, two different situations are considered: (1) all the states including the Inertial Measurement Unit (IMU) biases are estimated; (2) only the attitude, altitude, and vertical velocity are estimated. The second case is added as a safety feature to provide enough feedback signals to stabilize and land the quadrotor in the event of a position measurement loss, e.g. from a GPS due to jamming. A double loop control structure is implemented using an LQR controller with integral action, the inner loop contains the attitude and the altitude control, and the outer loop consists of x and y translational positions control. Finally, some preliminary results on the integration of C codes with Simulink using C MEX S-functions is described. A C library of a laser rangefinder sensor is transferred to a C MEX S-function to generate a 2D map of the environment using the laser sensor distance measurements to identify obstacles present within the range of the sensor. The concept of multi-threading and the integration of pthread library with Simulink using C MEX S-function are also described