A Deep Learning Based Device Authentication Scheme Using Channel State Information

Wide range implementations of Wi-Fi infrastructure in indoor environment facilitate the deployment of an autonomous device authentication system viable in field of network security, intruder detection etc. Existing device authentication systems standing in the literature suffer either from complicated key management or structural dependency. In this paper, we exploit a deep learning approach, namely Long Short Term Memory (LSTM) to authenticate devices more accurately utilizing Channel State Information (CSI) from received wireless signals. The proposed model doesn't require all CSI values to be stored and it increases the authentication accuracy significantly compared to the state-of-the-art works, as depicted in testbed implementation results.