A major challenge in mobile robotics is integration of methods into operational autonomous systems. Construction of such systems requires use of methods from perception, control engineering, software engineering, mathematical modelling, and artificial intelligence. In this paper it is described how such a variety of methods have been integrated to provide an autonomous service robot system that can carry out fetch-and-carry type missions. The system integrates sensory information from sonars, laser scanner, and computer vision to allow navigation and human-computer interaction through the use of a hybrid deliberative architecture. The paper presents the underlying objectives, the underlying architecture, and the needed behaviours. Throughout the paper, examples from real-world evaluation of the system are presented.