Self-organized criticality and stochastic learning based intrusion detection system for wireless sensor networks

In a constant search for efficient security control and intrusion detection systems (IDS), the ultimate goal in designing protocols remains less resource consumption while possessing broad coverage and wider applicability. Wireless sensors have become an excellent tool for military applications involving intrusion detection, perimeter monitoring, information gathering and smart logistics support in an unknown deployed area. Since sensor networks are resource-constrained devices, their design needs to minimize efforts without compromising the task's integrity. For this purpose, we propose a novel approach for an intrusion detection based on the structure of naturally occurring events. With the acquired knowledge distilled from the self-organized criticality aspect of the deployment region, we apply a hidden Markov model. In other words, we let our sensor network adapt to the norm of the dynamics in its natural surroundings so that any unusual activities can be singled out. Our IDS is simple to employ, requires minimal processing and data storage. Experimental scenarios are presented as verification of its functionality and practicality.