Volume 18, No. 6, 2021
Deep Learning Model For Predicting Context Information For Authentication
Amit Jaykumar Chinchawade , Onkar Singh Lamba
Abstract
Authenticating devices in a network with a large number of them is a difficult task. A huge number of data generating and data collecting devices may be connected in the internet of things (IoE). The authentication of each device is critical in a network with so many devices. The user equipment can be connected to the network via a fixed cable or wireless connection. The challenge of identifying such equipment or devices is difficult. Based on device address authentication is a crucial stage in every form of network that is practically familiar to every attacker approach and thus susceptible to being duplicated. With the periodic authentication system suggested in this research, duplication of devices that can steal crucial data can be avoided. This study presents a strategy based on the context of the user's equipment. The distance between the device and the connecting media affects the device's channel impulse response. The device's physical location and channel impulse response can be mathematically modelled and used as authentication context information. The experimental setup with deep learning-based channel impulse response prediction achieves excellent results in identifying the original device in a network emulation tool.
Pages: 1742-1750
Keywords: IoE, Device authentication, Deep learning, Context information, channel impulse response.