Deep space exploration is in the agenda of the major space agencies worldwide; certainly the European Space Agency (SMART Program) and the American NASA (New Millennium Program) have set up programs to allow the development and the demonstration of technologies that can reduce the risks and the cost of deep space missions. From past experience, it appears that navigation is the Achilles heel of deep space missions. Performed on ground, this imposes considerable constraints on the entire system and limits operations. This makes it is very expensive to execute, especially when the mission lasts several years and, furthermore, it is not failure tolerant. Nevertheless, to date, ground navigation has been the only viable solution. The technology breakthrough of advanced star trackers, like the advanced stellar compass (ASC), might change this situation. Indeed, exploiting the capabilities of this instrument, the authors have devised a method to determine the orbit of a spacecraft autonomously, onboard, and without a priori knowledge of any kind. The solution is robust and fast. This paper presents the preliminary performance obtained during the ground testing in August 2002 at the Mauna Kea Observatories. The main goals were: (1) to assess the robustness of the method in solving autonomously, onboard, the position lost-in-space problem; (2) to assess the preliminary accuracy achievable with a single planet and a single observation; (3) to verify the autonomous navigation (AutoNav) module could be implemented into an ASC without degrading the attitude measurements; and (4) to identify the areas of development and consolidation. The results obtained are very encouraging.