A cosine similarity-based compensation strategy for RSS detection variance in indoor localization

Currently, all kinds of Location-based Services (LBS) are gradually demanding more location information of mobile users. The WiFi-based indoor positioning technologies have been investigated intensely. However, both positioning accuracy and stability are often degraded by RSS detection variance between different devices. To solve the two problems, this paper proposes a compensation strategy for RSS detection variance based on cosine similarity. Specifically, it uses cosine similarity as the metric to determine whether different devices could conduct compensation for RSS variance. To obtain the device pairs that satisfy the metric standard, it uses ratio correction method to compensate for RSS detection variance thus effectively solves the positioning accuracy and stability reduction causing by RSS detection variance between different devices. Experiments have shown that the proposed strategy reduces positioning error and improves positioning stability significantly.