An adaptive virtual relaying set scheme for loss-and-delay sensitive WSNs

In loss-and-delay sensitive wireless sensor networks (WSNs), especially when the duty cycles of nodes are extremely low, it is a challenge to ensure that data can be transmitted to sink nodes with high reliability and low delay. To address this problem, in this paper, we propose a data collection scheme named Adaptive Virtual Relaying Set (AVRS) where a set of relay nodes with more reliable connections to the sender node is selected to form its Virtual Relaying Set (VRS) to help transmit packets. In ring-based WSNs, each node in a VRS helps send packets in turn to the upper ring before the transmission is successful, or the packet is dropped if they all fail. Therefore, the larger the VRS (implying more retransmission chances), the higher the packet transmission reliability and the lower the delay will be. On the other hand, as the sender node has to stay active during the transmission stage, having a large VRS will cause huge energy consumption. Combined with the fact that the energy consumption of different parts of WSNs is unbalanced, nodes in the near-sink area (i.e., hotspots) have extremely high energy cost while nodes in the far-sink area (i.e., non-hotspots) still have ample energy remaining when the network dies. The main idea of the AVRS is the following rule. The size of the VRS of nodes is determined adaptively according to its energy usage pattern, making the size small in hotspots and relatively large in non-hotspots. The AVRS takes advantage of the residual energy of nodes in far-sink areas to achieve improved network data collection performance. Meanwhile, as nodes in near-sink areas have small VRS, the network can maintain a long lifespan without any reductions. Both theoretical analyses and simulative results demonstrate that the AVRS can improve data transmission reliability by more than 50% and reduce network delay by at least 33%.