Machine Learning Based Access Point Verification Scheme for the Smart Grid

In the smart grid, there is increasing demand for monitoring and metering by using advanced wireless technology for the sake of massive connection, high transmission rate, and low latency. However, it is vulnerable to various attacks, such as rouge access point (AP), in wireless networks. In this paper, we propose an AP verification scheme to verify credibility of the target AP based on the channel state information (CSI). The XGBoost method is utilized to train the AP verification model offline. When the online verification process begins, the AP verification process is performed based on the CSI. Simulation shows that the proposed AP verification scheme achieves excellent performance of detecting rogue APs. Meanwhile, simulation results also show that the proposed AP verification scheme can resist rogue AP attacks.