Maria Charalampidou - Democritus University of Thrace
George Pavlidis - Institute for Language and Speech Processing, Institute / &#۰۳۹;Athena&#۰۳۹; Research Center
Spyridon Mouroutsos - Democritus University of Thrace

The necessity to tackle the increasing common concern about safety issues, urges the scientific community to come up with the development of innovative intruder detection and early warning security systems. One of the most effective technological solutions is provided by the application of WSNs. In this endeavor, most solutions have already adopted supercomputers and other computer resource systems to process the enormous amount of data. Alternatively to this approach, simpler and more easily implementable solutions, such as the WSNmod method (previously published by our research group), are already being put to use. In particular, WSNmod is based on three key elements, the categorization of sensor inputs, the quantization of the inputs and a time-window processing. WSNmod was introduced as an advanced intrusion detection system that focused on the minimization of the false positive alerts. Building on the idea of WSNmod, in this paper we focus, identify and quantify measurable parameters that influence the detection reliability. In addition, the very promising test results of the method and the security system are presented in a range of environmental conditions.

Wireless Sensor Networks (WSN), Sensor Selection, Ultrasonic Sensor, Motion Sensor, Vibration Sensor, Security, Intruder Detection, decision making, Sensors, sensor networks, Sensor Nodes, Policies, Security Applications, Surveillance Systems