Intelligent Operators for Localisation of Dynamic Smart Dust Networks

Wireless Sensor Networks (WSN) of Smart dust motes are becoming increasingly effective in environmental monitoring applications. In some applications, data gathered via WSN are only useful when combined with individual mote positions and time-stamps; if the motes are not static, it is important to find methods for their 3D location estimation from available RSSI signal data. Usually termed 'Localisation', this problem is made more complex when the positions of the motes are subject to external forces. Here we extend our previous work on solving this problem with evolutionary algorithms; we experiment with 'geometric-awareness' operators that constrain the positions a WSN mote can occupy to those that have a better probability of maximally contributing to the chromosome fitness. A hybrid comprising the specialised and standard operators is also tested. We conclude that this research direction offers a viable basis for a heuristically directed evolutionary system capable of suitably accurate 3D localisation, thus increasing the possibilities for viable WSN applications.