Machine Learning-Aided Indoor Positioning Based on Unified Fingerprints of Wi-Fi and BLE

This paper deals with an indoor positioning with the aid of machine learning based on the received signal strength indication (RSSI) fingerprints of beacon signals of both Wi-Fi and Bluetooth low energy (BLE). In fingerprint positioning, a site-survey is conducted in advance to build the radio map which can be used to match radio signatures with specific locations. Thus, it can take the impacts of empirical indoor environments into consideration. However, even if the physical positional relationship in the indoor environment is static, the observed RSSI values are dynamically fluctuated according to the probabilistic wireless channels. Unfortunately, it is difficult to analytically capture the stochastic behavior of RSSI in real-environments, and the accuracy of position estimation is degraded due to the model errors. To tackle this challenging problem, machine learning-based logistic regression is applied to fingerprint positioning with the RSSI data set (available as big data). Additionally, by exploiting a unified fingerprint generated from both Wi-Fi and BLE beacon signals, further performance improvement in the estimation accuracy is possible, owing to the transmit diversity effects. The experimental results show the validity of the proposed positioning scheme with the unified Wi-Fi and BLE fingerprint.