Is FM a RF-Based Positioning Solution in a Metropolitan-Scale Environment? A Probabilistic Approach With Radio Measurements Analysis

Recent mobile devices have already contained a low-cost FM receiving function due to the continuing improvements in the device manufacturing. This paper shows that positioning based on FM signal is an alternative radio option while meeting the FCC requirement. We present a probabilistic location system using a wider-covered and longer-lived FM infrastructure. The performance is evaluated in two different metropolitan-scale environments including National Taiwan University (NTU) and Wen-Shan rural area. Both results show that the FM based location system not only satisfies the FCC requirement but also provides a comparable or even better performance to GSM based solution. Moreover, we completely analyze the realistic radio measurements of FM and GSM from four perspectives including temporal variation, spatial separation, measurement correlation and spectrum allocation. Most FM measurements are observed to provide a lower temporal variation but a weaker spatial separation than GSM. Fortunately, we discover that the lack of spatial separation can be compensated by adding additional sensed channels. This property is useful especially for a rural area where the available GSM base stations are limited and distant. Furthermore, we point out that the spatial separation of GSM signal decreases when the signal level is weaker than -90 dBm. At such a condition, FM reports a better accuracy than GSM even with the fewer channels.