A Deep Learning Based Bluetooth Indoor Localization Algorithm by RSSI and AOA Feature Fusion
Dekang Zhu, Jun Yan
Abstract
With the development of the Internet of Things, location based services has received much attentions. Bluetooth 5.1 standard provides the Angle of Arrival (AOA) direction finding function, which opens a new approach for indoor Bluetooth localization. In this paper, a deep learning based Bluetooth indoor localization algorithm by received signal strength indicator (RSSI) and AOA feature fusion is proposed. For data preprocessing, the principal component analysis (PCA) is used to reduce the redundancy of RSSI measurement. The Kalman filter is used to smooth AOA measurement. Then, a convolutional neural network (CNN) is used for feature extraction which extracts the deep-level features of RSSI and AOA measurement respectively. After feature fusion for the above two features by concatenating operation, the Softmax layer is used for classification learning. At last, the localization classification model is obtained. The experimental results show that, compared with the existing localization algorithms, the proposed algorithm has significantly improved the localization performance.