Accuracy Evaluation of Indoor Positioning by Received Signal Strength using Deep Learning

In this paper, we examined indoor positioning systems that combine the deep learning technology with the fingerprint using the received signal strength indicator(RSSI) of Wi-Fi. Because the fingerprint method used previously recorded data, positioning can be performed considering effects in actual indoor environments to obtain a high-precision result compared to other methods that use theoretical formulas. The accuracy of deep learning depends on data shaping and learning methods. Therefore, this study aimed to compare existing methods' accuracy by determining compatible shaping and learning methods. The effectiveness of the proposed method was demonstrated by comparing it with the existing methods.