A Survey on Detection of Malicious Nodes in Wireless Sensor Networks

Abstract

Wireless Sensor Networks (WSN) are often utilized in different places where human access is limited to monitor physical and environmental factors. They are vulnerable to errors and malicious attacks because of low resources and use in hostile environments. The impacted or impaired sensor nodes can deliver misleading or wrong data to the base station. The malicious nodes in the network will intentionally drops the data packets, modifies the data contents and delay the data transmission that reduces the system performance. Therefore, it is necessary to accurately and promptly identify malicious and malfunctioning nodes to ensure that the networks function reliably. In this article, numerous malicious node detection models in wireless sensor networks are discussed and the process is clearly illustrated to identify the drawbacks. In the presence of natural defects and noise, malignant nodes are successfully recognized without surrendering defective nodes. In order to prevent faked information injecting by the adversary via hacked nodes, it is necessary to detect and isolate compromised nodes. However, because of low scalability and high overhead communication, it is difficult to protect flat topology networks efficiently. A brief survey is presented on different trust-based models aimed at WSNs for malicious node detection. Malicious node detection causing attacks, packet loss causes, data modifications are the challenges to be overcome due to malicious nodes in the network. Moreover, different sorts of malicious attacks on trust model are identified and whether the existing trust models can withstand these attacks or not has been assessed.