WSN Network Node Malicious Intrusion Detection Method Based on Reputation Score

Abstract

Wireless sensor network (WSN) is the Internet of Things by a large number of sensors in the external physical environment to obtain data information, and use wireless communication technology to provide users with information transmission services. At this stage, communication and security mechanisms are the main problems faced by WSN. This is because most of the existing sensors are powered by batteries with very limited energy, and most of them are deployed in an outdoor open environment, which is easy to be captured as a malicious node. Network attacks. However, the existing malicious node detection methods have shortcomings such as low efficiency, high energy consumption, and insufficient performance. Therefore, this paper proposes a WSN malicious node intrusion detection method based on genetic algorithm optimization of LEACH hierarchical routing protocol. Based on the optimization of the LEACH protocol by genetic algorithm, the method integrates the reputation evaluation mechanism, and screens and eliminates malicious nodes by calculating direct reputation, indirect reputation and comprehensive reputation, thereby ensuring the safe operation of WSN. The simulation results show that this method can effectively resist the attack of malicious nodes on WSN, and has obvious advantages over other methods.

A Review of Security Challenges and Solutions in Wireless Sensor Networks

“…WSNs are the backbone component of the IoT, which is widely used for universal connectivity, security, control, and general awareness of things. Low-powered, small-sized sensors can efficiently monitor and collect data from any environment [1,2,14,20,21,22,23]. "Things" in the IoT refer to tiny embedded physical sensing devices (i.e., sensor nodes) that are connected to the Internet to achieve a specified task [4].…”

“…WSNs provide special challenges for security protocol designers due to the following unique characteristics [6,7,8,15,19,21,22,23,26,36,37,39,40,41]:…”

“…The installation and initialization of an IDS with pre-knowledge before deployment is impractical, and post-deployment learning and intrusion detection activities are expensive in terms of computing overhead and energy consumption. Therefore, current IDSs are not suitable for wireless sensor networks [13,22,23,36,37].…”

“…Here are some approaches for intrusion detection using ML in WSNs: Support Vector Machines (SVMs), k-Nearest Neighbor algorithm (KNN), hybrid detection, reinforcement learning, and transfer learning. However, the problem remains with the machine learning training process, so many studies have attempted to improve it by decreasing training time, depending on a small data set, and enhancing accuracy [1,5,14,15,22,23,37,44,53].…”

Construction and Analysis of QPSO-LSTM Model in Network Security Situation Prediction

“…For network security situational awareness technology, the acquisition of situational elements, situational assessment, and situational prediction are important key links in the research field of this technology. Network security situational awareness technology can integrate internal and external network environment information, and conduct comprehensive analysis and evaluation of the collected relevant information [5]. By conducting situational assessment on the results obtained from fusion calculations, it is possible to provide a more comprehensive analysis of network security situational information while ensuring accuracy.…”