RSSI Purification Model Based on Time and Frequency Domain Filtering

10.29428/9789860544169.201801.0002

Chinyang Henry Tseng；Shin-Fu Tsai

Indoor positioning ； Received Signal Strength Indicator ； Gaussian filter ； Fourier transform

NCS 2017 全國計算機會議

2017（2018 / 01 / 01）

7 - 11

英文

As wireless network and mobile devices evolve rapidly, indoor positioning has become a critical service. Our daily environment is filled with wireless signals, which can deliver the indoor positioning service. To improve the quality of indoor positioning, this paper proposes the purification of RSSI (Received Signal Strength Indicator), which increases the accuracy of indoor positioning. To acquire the appropriate RSSI value from the RSSI sets, the proposed purification model applies Gaussian filter to abstract the frequency and time domains of the collected RSSI sets. The best abstraction order is frequency, time, and frequency. In addition, the abstraction of the frequency requires Fourier transform together with Gaussian filter to deal with the complex array of RSSI values. Then we implemented the model on Broadcom Bluetooth chips. The experiment results show that the proposed model has better distance errors than traditional Gaussian filter, which only applies the purification on time domain.