Quantifying importance of reference locations for RSS-based localization

Collecting onsite received signal strength (RSS) is commonly employed in RSS-based positioning systems to achieve more accurate estimations. However, it requires a time-consuming site survey and a great amount of manual efforts. This study proposes a novel approach to reduce the need of massive RSS collection by quantifying the importance of each RSS-measured location. The proposed approach, namely minima inverse linearity (MIL), investigates the linearity in local regions by considering the pairwise power-distance between base stations. MIL assumes that the inverse-linear regions may provide more information since the behavior is against the radio propagation model. It gives higher scores to those reference locations while quantifying the importance. After ranking all reference locations using the quantified importance, we can select better measured-locations to configure positioning systems and reduce the massive RSS collection. This algorithm is useful when the reference locations are limited due to constrained resources. The experiments applied MIL to an actual GSM network. The results show the effectiveness of MIL. Given limited measured-locations, the proposed algorithm can efficiently suggest important reference locations, indicating where we should measure RSS first to enhance the positioning system.