| Abstract: |
Wireless Sensor Networks (WSNs) experience increasing capabilities of sensor nodes challenging operational safety and security, which results in a new research focus on the resilience of sensor nodes. Anomaly Detection Systems (ADSs) on the other hand are a well-established means of identification of abnormal behavior or data. Consequently, the main research challenge of this Ph.D. thesis lies in the adoption and adaptation of ADSs in WSNs with their specific characteristics and limitations such as battery life, spatially dispersed distribution, exposition to harsh environmental conditions, and attacks on network and physical level.
The specific outcome of the Ph.D. thesis shall comprise (i) an extensive evaluation of available detection schemes, (ii) guidelines and best practices for applying anomaly detection in WSNs, and (iii) a distributed anomaly detection approach specific to hierarchical WSNs. The novelty of this research stems from (i) the adaptation of well-known ADSs to the new boundary conditions of WSNs and (ii) the development of new concepts for effective, lightweight and decentralized ADSs. |