Personnel Fingerprint Positioning Algorithm Based on SVM Classification in Underground Coal Mine

In order to reduce influence of underground environment on the positioning system, this paper presented a fingerprint positioning algorithm based on SV M classification. The algorithm consists of fingerprint database, underground tunnel fingerprint data acquisition and underground location matching. The SVM classificat ion method was adopted to establish the fingerprint database, the singular value was removed to eliminate the effects of dynamic fingerprints, the best match position w as found by matching the real- time sampling signal to the fingerprint database. This paper collected 50 samples by randomly fingerprint data as position information, m uti- terminal user position information was measured and five user terminals measurement data were taken for analysis. The results of location experiments showed tha t the positioning error of proposed algorithm was less than 1. 5 m, which was much better than traditional positioning algorithm based on RSSI.