The conventional clustering method has the unique potential to be the framework for power-conserving ad hoc networks. In this environment, studies on energy-efficient strategies such as sleeping mode and redirection have been reported, and recently some have even been adopted by some standards like Bluetooth and IEEE 802.11. However, consider wireless sensor networks. The devices employed are power-limited in nature, introducing the conventional clustering approach to the sensor networks provides a unique challenge due to the fact that cluster-heads, which are communication centers by default, tend to be heavily utilized and thus drained of their battery power rapidly. In this paper, we introduce a re-clustering strategy and a power-limit constraint for cluster-based wireless sensor networks in order to address the power-conserving issues in such networks, while maintaining the merits of a clustering approach. Based on a practical energy model, simulation results show that the improved clustering method can achieve a lifetime nearly 3 times that of a conventional one.

In this letter, we demonstrate a Relay-based Adaptive Auto Rate (RAAR) protocol that can find a suitable relay node for data transmission between transmitter and receiver, and can dynamically adjust its modulation scheme to achieve the maximal throughput of a node according to the transmission distance and the channel condition. Evaluation results show that this scheme provides significant throughput improvement for nodes located at the fringe of the AP's transmission range, thus remarkably improving overall system performance.