An Efficient Indoor Event Detection Mechanism Using Wireless Sensor Network

Wireless Sensor Networks (WSNs) have been widely implemented for environmental monitoring and event detection scenarios. Several existing works on event detection used the static or threshold values to assert whether an event has occurred or not, which led to imprecise sensor readings. A recently conducted work has used the fuzzy logic to manage the fluctuating readings of sensors and reduce the number of generated false alarms. However, several issues should be addressed such as in heterogeneous, different sensor types can affect the efficiency and accuracy of the fuzzy membership function and increase the number of inputs, which can increase the complexity of the fuzzy logic process. In this work, a new mechanism called probabilistic collaborative event detection (PCED), which is a hybrid event detection mechanism, is proposed based on the proposed clustering WSN topology, where an adopted probabilistic approach for heterogeneous sensor nodes is used to convert the sensing values to a probability form. Accordingly, the cluster head decision mechanism is proposed. At the fusion centre level, the fuzzy logic is introduced to enhance the accuracy of detection. The effectiveness of the proposed method was evaluated by using MATLAB software, which showed high detection probabilities in addition to low false alarm probabilities. PCED is compared with the well-known existing event detection mechanisms (such as the REFD mechanism). Thus, it was found that PCED has minimized the number of false alarms from 37 to 3 in some scenarios. PCED has also increased accuracy up to 19.4% in comparison with the REDF mechanism and achieved a reduction of detection latency up to 17.5%.