In this paper we describe a control paradigm for the navigation of a mobile robot toward a moving landmark in an unknown indoor environment with obstacles. The overall control paradigm has been successfully tested on a real integrated navigation system which comprises a webcam and a mobile platform equipped with odometric, sonar and bumper sensors. The webcam is used for landmark recognition while sonar, bumper and odometric sensors are used for navigation purposes. In the paper we present a hardware description of the considered experimental system and discuss the proposed control algorithms. Guidelines regarding both hardware and control algorithms are drawn to improve the efficiency of the overall system.