Abstract v1,3

This thesis deals with trajectory tracking and obstacle avoidance of autonomous multicopters. The goal is to design and implement a model predictive controller on an embedded platform for fully autonomous flight. We present a method of transforming a trajectory tracking problem into a quadratic optimization task. The obstacles, detected by onboard cameras, further linearly constrain the optimization task. A solver for linearly constrained quadratic programming has been designed and implemented on an embedded platform. The conducted real flight experiments verified the capability of the controller to stabilize the quadcopter, track trajectory and avoid obstacles.