Indoor positioning has gained more and more interest lately since it offers the possibility of using the location information in several network functionalities improvement or user services deployment. In this paper, we consider the case of WLAN-based indoor positioning for its simplicity and ease of deployment. With our focus on scene analysis methods, we study both deterministic and probabilistic schemes and try to improve their accuracy performance without increasing the processing and hardware requirements of the system. More precisely, orientation-based sampling of WiFi signal strength (SS) measurements and simple pre-processing of these samples are proposed for mitigating the inherent impairments of the wireless medium. Results based on real experimental measurements showed a considerable accuracy increase especially for the deterministic case