Volume 10, Number 6

Improvement of False Report Detection Performance Based on Invalid Data
Detection Using Neural Network in WSN


Sanghyeok Lim and Taeho Cho, Sungkyunkwan University, Republic of Korea


WSN consists of a number of nodes and base stations and is used for event monitoring in various fields such as war situations, forest fires, and home networks. WSN sensor nodes are placed in fields that are difficult for users to manage. It is therefore vulnerable to attackers, and attackers can use false nodes or MAC injection attacks through the hijacked nodes to reduce the lifetime of the network or trigger false alarms. In order to prevent such attacks, several security protocols have been proposed, and all of them have been subjected to MAC-dependent validation, making it impossible to defend against false report attacks in extreme attack circumstances. As attacks have recently become more diverse and more intelligent, WSNs require intelligent methods of security. Based on the report information gathered from the base station, the proposed method provides a technique to prevent attacks that may occur where all MAC information is damaged by carrying out verification of a false report attack through the machine learning based prediction model and the evaluation function.


Network Protocols, Wireless Sensor Network, simulation, machine learning, neural network