A Model Predictive Approach to Fault-Tolerant WASNs

In wireless ad hoc sensor networks (WASNs), a crucial issue is to reduce power consumption while satisfying some key network properties. In this paper, we propose a fault-tolerant topology control that optimizes the lifetime of the network at a given degree k of connectivity by minimizing power consumption. Our topology control is fully distributed and uses a model-based transmission power adaptation strategy based on model-predictive control. Specifically, the future network behavior is predicted in order to derive an optimal transmission power assignment which tracks the desired connectivity level minimizing energy. Our experimental results show that our localized solution is more scalable and requires much less communication bandwidth and energy than the centralized approach.