A eural etwork approach for Wireless sensor network power management

In this paper we introduce a new approach for wireless sensor network power management which is based on eural etworks. In this new approach an intelligent analysis is used to process the structure of a wireless sensor network (WS ) and produce some information which can be used to improve the performance of WS s’ management application. We applied our intelligent method to our previously proposed management approach which uses the concept of Multi-Agent systems for WS s’ management and observed the improvement of the performance. Wireless sensor networks need to be managed in different ways; e.g. power consumption of each sensor, efficient data routing without redundancy, sensing and data sending interval control, etc. The random distribution of wireless sensors, numerous variables which affect WS ’s operation and the uncertainty of different algorithms (such as sensors’ self-localization) give a fuzzy nature to WS s. Considering this fuzzy nature and numerous details, a neural network is an ideal tool to be used to cover these details which are so hard to be explicitly discovered and modeled. In this paper we introduce our eural etworkbased approach which results in a more efficient routing path discovery and sensor power management. We define a set of attributes based on sensors’ location and neighborhood and we use them as inputs of our neural network and the output of the neural network will be used as a factor in the route path discovery and power management. We designed a simulator based on our approach and observed the effect of our method on wireless sensor network lifetime and sensor power consumption which will be presented in this