In recent years, more and more robotic systems have been supporting rescue forces in their missions. This paper presents the control algorithm and its application of an unmanned areal vehicle (UAV), which can support emergency personnel in their work. An adaption of the so-called Limit Ellipsis Controller (LEC) for indoor UAV is proposed. This adaptation enables the LEC to be used in semi-structured environments with static and dynamic obstacles. The main benefit of the LEC is that it can prevent deadlocks caused by other methods in complex environments. Furthermore, the LEC function is implemented on an experimental UAV system and tested in various environments. The proposed technical system enables exploration of a building by a semi-autonomous UAV, saving valuable rescue time. The UAV can fly into the building and explore the interior without collision. The results show that the proposed controller can adequately avoid local minima, guide the UAV to the desired target, and provide essential information for the rescue team in real demonstration scenarios.