Common localization approaches are usually based on omni-directional antennas. There are various algorithmic methods, all of which assume strictly uniform radiation in all directions. However, in general, the output characteristics of these (nominally omnidirectional) antennas are not completely omnidirectional or even highly directional depending on the specific design. This article investigates the use of clustering algorithms to identify different areas of joint signal parameters that can be used for an RSS-based localization method. A set of clustering algorithms is investigated, and their properties to cluster RSS maps for localization applications are compared. In particular, their score for cluster quality and their execution speed relevant for deployment in embedded systems has been evaluated.