User Authentication Using Neural Network in Smart Home Networks

Security has been an important issue in the smart home applications. In home networks with distributed architectures that consist of a broad range of wired or wireless devices, it is likely that unauthorized access to some restricted data or devices may occur. Therefore, it becomes important to consider issues of security, authentication and access control. The authentication and authorization of users in smart environments are the key factors in the security of home networks. User authentication has been traditionally based on PIN, password, key, smart card or biometrics. Password-based authentication is widely used to identify legitimate users, because passwords are cheap, easy and reasonably accurate. In conventional password-based authentication methods, passwords store as a password or verification table. These methods use some encryption algorithms to prevent the passwords from being revealed, but they are still vulnerable. In this paper, we train a neural network to store encrypted passwords and use it instead of the password or verification table. This proposed method can solve the security problems in some authentication system and can be used to store the user profiles and access controls in smart home networks.