Research On Mobile Access And Optimization Of Information Acquisition In Wireless Sensor Networks

Some important characteristics of large-scale heterogeneous wireless sensor networks such as throughput, lifetime, reliability and coverage can be greatly improved by introducing mobility. However, fast and seamless mobile access can not been fully supported by current sensor network protocols, which results in the degradation of information acquisition in mobility enabled sensor networks. To explore large-scale heterogeneous and mobile sensor networks, this research focuses on three major aspects: mobile access, mobility centric topology optimization and mobility adaptive cross-layer optimization. The main contributions of this thesis are presented as follows.First, this thesis analysis some unresolved problems in mobile sensor networks. An protocol stack for heterogeneous mobile wireless sensor networks was proposed, which was based on the three-tired network architecture.Second, a mobile adaptive MAC protocol MA-MAC was introduced, which provides functionalities of low latency mobile access and automatic link maintenance. A transmission mode adjustment mechanism was also proposed to alleviate the funneling effect around mobile sinks. Simulation shows that compared to current mobile MACs, MA-MAC greatly improves the performance of data exchange between mobile sinks and static sensor nodes. Furthermore, it is simple and easy to implement.Third, by combining MA-MAC and the existing double-clustering based MAC protocol DC-MAC, a mobility enabled MAC protocol for clustered sensor networks MA-DC-MAC was proposed. In order to improve the performance of convergecasts toward mobile sinks, we introduced an optimization mechanism called AOCMSN (Adaptive Optimization for Clustered Mobile Sensor Networks). Simulation shows that the combination of MA-DC-MAC and AOCMSN significantly outperforms the original DC-MAC in throughput, delivery ratio, data transmission latency and energy consumption. Fourth, we took a deep investigation into the performance of wireless sensor networks with hybrid sinks and proposed an information acquisition method which took advantage of the hybrid network model in aspects like high success rate in data transmission and low energy consumption.Finally, we extended the Qualnet network simulator and provided a highly accurate and effective simulation platform for heterogeneous mobile wireless sensor networks. All the experiments in this thesis were carried out on this platform.