Adaptive Trust Management Protocol based on Intrusion Detection for Wireless Sensor Networks

Wireless Sensor Networks run critical applications and need to be protected against malicious attacks and faults. In this paper we propose Adaptive Trust Management Protocol, a protocol that adjusts trust and reputation based on node behavior. The protocol includes three phases: the Learning phase, in which experience is computed based on these alerts received from TinyAFD, the Exchanging phase, in which experience associations are exchanged between neighbor nodes, and the Updating phase, in which trust and reputation are updated based on experience. ATMP has been implemented on top of TinyOS and has been tested using TOSSIM in several attack scenarios to evaluate the evolution of experience, trust and reputation. We performed a comparative evaluation of ATMP with other similar solutions and determined that our protocol covers a larger range of attacks, mostly because of the integration with a complex intrusion detection system.