Metrics for Comparing Different Types of Positioning Techniques: Proximity Detection and Location Coordinate Estimation

Satoki Ogiso, Satoshi Nakae, Karimu Kato, Ikue Mori, Takahiro Miura, Ryosuke Ichikari, Takashi Okuma, Takeshi Kurata, Takanori Sugi, Yasunori Haga, Eri Nakano, Atsushi Kimura

Abstract: Indoor positioning methods can be categorized into two main types: proximity detection, which identifies the area where the target is located, and location coordinate estimation, which determines the coordinates of the target. When introducing a positioning system, it is sometimes necessary to compare these different output systems, and it is difficult to make quantitative comparisons. This paper proposed a set of metrics to compare outputs of proximity detection and location coordinate estimation. Experiments were conducted in a factory to compare proximity detection and location coordinate estimation in an environment similar to that in practical use. From the results, the precision, recall, and coordinate-based errors of the proximity detection system are slightly worse than those of the location coordinate estimation. Also, it was found that the proximity detection system was able to acquire about 40% of the data compared to the location coordinate estimation. Through this comparison, we confirmed the effectiveness of these metrics to show the strengths and weaknesses of compared systems.