A novel method for the preliminary identification of asteroids at discovery and a few days thereafter is being developed in Helsinki. Having two different sets of asteroid observations, the goal is to identify all possible pairs of objects between the sets. An arbitrary asteroid can either remain unidentified, or be preliminary linked to one or more asteroids. In the case of ambiguity, the final decision must usually be based on additional observations. We use a multistep approach, during which possible pairs of objects are first selected by comparing ephemerides that have been generated for three common epochs. The method has been successfully tested using both Very Large Telescope observations, and simulated observations of near-Earth and main-belt objects. Identification results of simulated observations indicate that the observing strategy promoted by the Minor Planet Center might not be the best one, at least for the purposes of identification. The ultimate goal is to produce a real-time asteroid identification tool for ESA's astrometric space observatory Gaia, the Lowell Observatory Near-Earth-Object Search, the Near-Earth Space Surveillance mission, and the Nordic Near-Earth Object Network. The tool could also benefit large-scale surveys done with the Large Synoptic Survey Telescope, and the Discovery Channel Telescope.To search for other articles by the author(s) go to: http://adsabs.harvard.edu/abstract_service.html