Performance Optimization of Reconfigurable Real-Time Wireless Sensor Networks

Abstract

Wireless sensor networks (WSNs) can be seriously impacted by several changes in behavior. As a measure of optimality, a network should react in real-time. However, even if some reactions can make the system flexible, they can cause significant damages when they are not well-structured. Moreover, the scenarios of reconfigurations can affect several parameters within a wireless sensor network. Among these parameters it is possible to name the energy efficiency, the memory limitation within a node, the real-time constraints at the level of the nodes and the network in general. Either through the application or the transmission of the reconfiguration scenarios, some parameters are severely altered. A metamorphosis of the internal architecture of the nodes is proposed in this research work as well as a policy of communication between several nodes aiming to adapt the network to any change. The resulting proposition offers general efficiency in energy and real-time constraints. The efficiency is realized by the application of a pipelined approach, dealing with the incoming reconfiguration scenarios, and a communication protocol, based on a priority energy and deadline aware scheduling algorithm (a multicriteria scheduling algorithm). This protocol is capable of optimizing the transmission of both reconfiguration scenarios and sensed data simultaneously. The solution is tested on a network connecting several cars during an automobile journey and the results validate the efficiency of this proposition.