Optimising energy in distributed phenomena detection for mobile WSN

We propose an energy-aware distributed scheme, general phenomena detection, to detect phenomena in data gathered from mobile sensors. In the proposed algorithm, mobile sensors self-organise themselves into groups and elect group heads (GHs) based on the location of the phenomena. To better share, the extra battery power overhead, GHs and subsequently group membership are updated periodically (every window). Each GH gathers readings from sensors within its group and processes the data to detect possible phenomena within its geographical boundaries. Then, GHs communicate with each other to discover and report global phenomena. To further reduce the energy cost, the study proposes three optimisation strategies: the first strategy, reporting by partial participation (RPP), limits the number of participating GHs in reporting the phenomena. RPP achieved a saving in the energy cost of around 50%. The second strategy, reporting by Z-order (RZO), provides an overall short communication path between GHs by using Z-order. RZO achieved a saving of more than 35% of the required energy. The study also proposes a lazy window update strategy that is suitable for a wireless sensor network (WSN) with slow sensor speed. The proposed solutions are validated via comprehensive experiments performed on an NS2 network simulator.