Topology control in lossy wireless sensor networks with delay constraint

Traditional topology control designs of wireless sensor networks (WSNs) usually assume reliable links between two sensors if their distance is less than a threshold. In real environments, however, relaxing the reliable threshold to utilize lossy links is beneficial for saving node transmission power so that the network IifecycIe can be prolonged. Since transmission over lossy links is unreliable, data retransmission is required. But retransmission results in extra energy consumption and remarkable communication delay. In view of this, we define an enhanced version of restricted shortest path (RSP) problem, called end-to-end delay-constrained topology control problem in lossy WSNs (Lossy-RSP), which takes into account retransmission costs in terms of energy and time. The difficulty in the Lossy-RSP problem is to deal with parallel edges collision phenomenon, i.e., different power levels between two adjacent nodes are suited for constructing constrained shortest paths between different sources and destinations. The Lossy-RSP problem is NP-complete in strong sense due to parallel edges collision phenomenon. We present two independent heuristic algorithms to defuse this phenomenon while retaining energy-efficient links. Simulation results show that our algorithms can reduce the energy consumption significantly even under rather strict delay constraint.