Length-aware topology reconfiguration in wireless sensor-actor networks to recover from an actor failure

We consider the impact of actors' failure on the connectivity in wireless sensor and actor networks. Most of the previous works focus on the failure of cut-vertexes, which may cause the network partitioned into disjoint components. However, since the failure always happens abruptly and randomly, a particular set of actors may exists whose failure would prolong the link length between any pair of others observably. This paper regards actor failure as the impact on link length between actors, and presents a novel parameter named as length Impact Index (LII) to estimate significance of each actor to the whole network. Based on the LII, the Length-aware Topology Reconfiguration Algorithm (LTRA) is proposed. LTRA is a distributed and self-healing algorithm, which relies on the local topological information to devise a reconfiguration plan that can relocate the least number of actors. The performance of LTRA is validated by extensive simulation experiments.