Trust management is an effective technique for dealing with malicious and compromised nodes in Wireless Sensor Networks (WSNs). It has demonstrated its benefits in various issues such as secure cluster head (CH) election, secure localization, secure routing and misbehavior detection. This work presents a secure CH election algorithm and a misbehavior detection approach. Multiple metrics were used for the CH election, including the key metric for the election which is the trust degree of the sensor node. The problem of selecting the most trustworthy node as CH was also addressed. In addition, a monitoring strategy to evaluate the behavior of sensor nodes using multiple trust types was developed. Thus, the aim was to keep only the trustworthy nodes in the network and eliminate the malicious nodes. The case of a compromised CH was considered; a trust evaluation mechanism at the cluster members level and a local clustering algorithm were adopted to isolate the malicious CH without affecting the network performance. The simulation results indicate that the proposed scheme prevents the malicious nodes from becoming CHs and protects the network from compromised CH after the election. With respect to the misbehavior detection, the proposed scheme achieved a high detection rate of malicious nodes with a low number of false positive and false negative alarms.

信任管理是一种有效的技术，可以处理无线传感器网络（WSN）中的恶意节点和受感染的节点。它已在各种问题上证明了自己的优势，例如安全的簇头（CH）选举，安全的本地化，安全的路由和不良行为检测。这项工作提出了一种安全的CH选择算法和不良行为检测方法。CH选举使用了多个度量，包括选举的关键度量（即传感器节点的信任度）。还解决了选择最值得信赖的节点作为CH的问题。此外，开发了一种监视策略，用于使用多种信任类型评估传感器节点的行为。因此，目的是仅将可信任节点保留在网络中，并消除恶意节点。考虑了CH受损的情况；采用集群成员级别的信任评估机制和局部集群算法隔离恶意CH，而不影响网络性能。仿真结果表明，该方案可以防止恶意节点成为CH，并在选举后保护网络免受CH的侵害。对于不当行为检测，该方案实现了对恶意节点的高检测率，误报和误报警报数量少。