Title: A congestion attack behaviour recognition method for wireless sensor networks based on a decision tree
Authors: Wen Feng; Xuefeng Ding
Addresses: Informatization Construction and Management Office, Sichuan University, Chengdu, 610065, China ' Informatization Construction and Management Office, Sichuan University, Chengdu, 610065, China
Abstract: To overcome the problems of low attack recognition rates and high congestion identification errors in traditional congestion attack behaviour judgement methods for wireless sensor networks (WSNs), a new congestion attack behaviour judgement method based on decision trees is proposed in this paper. First, the traffic is processed, a congestion attack detection model is constructed as a decision tree, and the sensor nodes with artificial labels are classified and evaluated. According to the classification results, the network congestion attack behaviour is identified according to an attack recognition threshold. The experimental results show that the average recognition rate of congestion attacks is 99.82%, the error rate of congestion identification is only 0.00284, and the packet loss rate is low, indicating that this method can effectively recognise network congestion attack behaviours.
Keywords: low attack recognition rates; high congestion identification errors; decision tree; WSNs; congestion attack; attack behaviour judgement.
International Journal of Sensor Networks, 2021 Vol.36 No.4, pp.236 - 242
Received: 22 Jan 2021
Accepted: 22 Jan 2021
Published online: 21 Aug 2021 *