Accurate wi-fi signal strength recovery method using chebyshev wavelet- based approximation for indoor positioning

In many scenarios of everyday life and especially in warehousing, manufacturing and logistics, it is highly desirable to locate objects or persons quickly and accurately. Nowadays, fingerprinting based Wi-Fi positioning system provides enterprises the ability to track their various resources more efficiently and effectively. The main idea behind fingerprinting is to build signal strength database of target area prior to location estimation. This process is called calibration and the positioning accuracy highly depends on calibration intensity. Unfortunately, calibration procedure requires a huge amount of time and effort and makes large-scale deployments of Wi-Fi based indoor positioning systems non-trivial. In this research, we present a new method of recovering Wi-Fi Radio Map (WRM) database based on few sample received signal strength indicators –RSSI and this recovered data is used as radio map – constructed Wi-Fi RSS based fingerprint database for indoor positing. In contrary to conventional calibration method, our method requires only a few signal samples to be collected and rest of the data are approximated using Chebyshev wavelets. The main goal of our research is to minimize the calibration workload while maintaining recovered data accuracy and achieve acceptable results on positioning accuracy. Compared to the conventional way, proposed a new method to construct accurate Wi- Fi signal strength indicators using Chebyshev wavelet based approximation requires only a few reference RSSI samples, and this significantly will reduce the calibration effort. Also, field test results showed that proposed method achieves better approximation accuracy than existing interpolation methods, such as VORO and MOSM. Also recovered RSSI data - fingerprint database was used with positioning software to evaluate results of positioning accuracy. Results show that positioning accuracy is significantly improved compared with conventional, as well as, other two related methods.

Preview Text FSKTM 2016 16 IR.pdf Download (2MB) | Preview

Actions (login required)

View Item