基于曲面拟合的WiFi指纹数据库更新

WiFi fingerprint database updating based on surface fitting

针对指纹数据库更新需要大量时间和人力开销问题,提出了一种新的WiFi环境下接收信号强度(RSS)的估计方法。即采用径向基函数插值的方法,利用一部分RSS被重新测量的参考点,拟合出接收信号强度曲面,估计出邻近未知参考点RSS值,从而更新指纹数据库。大量实验证明,所提方法只需要测量四分之一的参考点,就能在定位误差为2 m范围内达到与实测数据库相同的累积误差概率,保证了较高定位精度。

To solve the problem that the fingerprint database updating requires a huge amount of time and laboring effort,a new Received Signal Strength（ RSS） estimation method was proposed in WiFi environment. Part of remeasured RSS fingerprints was used to fit RSS surface by the radial basis function interpolation,the RSSs of the unknown reference points were estimated nearby the known reference points,and then the whole fingerprint database was update. The extensive experiments prove that only a quarter of the reference points need to be remeasured,the cumulative error probability will be similar with the actual database within the positioning error of 2 m,and thus the method guarantees the satisfactory positioning accuracy.