Research On In-cluster Node Energy Manage-ment Methods In Wireless Sensor Networks

The lifetime of wireless sensor network (WSN) is restricted by the energy capacity of battery. Sensor nodes are generally powered by battery, whose capacity is very limited, and it is difficult or inconvenient to replace. Energy problem is the most critical problems in wireless sensor networks when design. In recent years, with the development of environmental energy harvesting technology, it’s an effective way to harvest ambient energy from the environment to realize self-powered for WSN nodes. But, this method can’t fundamentally solve the energy supply problems of WSN because of the limitation of the current standard of technology, cost and the uncertainty and randomness of energy harvesting. Therefore, to reduce power consumption from every details, and to realize low power consumption and optimized resources allocation through energy management, so as to meet the performance index of WSN, including quality of service and network life, becomes the main goal in the design of WSN.According to the architecture and characteristics of WSN, this thesis analyzes the energy consumption problem in WSN, and summarizes the node level and network level energy management methods in WSN. Based on the existing principles and methods of dynamic power management (DPM), this paper proposes a DPM algorithm based on exponention-average prediction strategy, which has better energy saving effect when it is comparing with two typical DPM algorithm (named as probability-based DPM algorithm and wavelet decomposition based DPM algorithm) in WSN. Inspired by the thought of clustering and collaborative work among sensor nodes, this thesis puts forward an in-cluster node energy management algorithm which can guarantee the field coverage and network connectivity. Performance comparison analysis of this algorithm such as energy consumption and the network’s life cycle is carried on, simulation experiments show that the proposed algorithm is effective. Finally, according to energy supply characteristics of the solar, this paper predicts the energy consumption and energy supply level of the sensor nodes, applies the energy management method proposed before into the situation the situation which the sensor nodes may abstract ambient energy from the environment. The simulation results show that this algorithm can adapt to the situation which the sensor nodes with energy supply, and has better energy balance.