Research On Key Technologies For Energy-Efficient Data Query In Wireless Sensor Networks

A wireless sensor network (WSN), which is reported as one of the most influential technologies in the 21st century, will take great impact on human future life. Holding the features of low-cost, easy-to-deploy, high-resilience and good-scalability, WSNs have tremendous applicable value and can be widely used in many application areas, such as military affairs, environmental protection, health cares, human life, disaster salvage and so on. Therefore, WSNs have attracted people's attention from both industry and academia. Different from the traditional wireless communication networks, WSNs exhibit many special features, such as data-centric, energy-constrained, topology-dynamic and application-related, which have brought many challenging problems for technologies on topology control, routing protocol, data query and application system. Aiming at achieving the energy efficiency, this dissertation studies the key technology issues of data query in WSNs, in which the problems of minimizing the energy consumption in transmission are discussed and the solutions are proposed.In WSNs, when a sink node need to extract the query result from all the raw data of multiple source nodes, the in-network data fusion technology can be adopted to reduce the amount of data transmitted, hence the energy consumption in transmission. As a result, how to choose the only one fusion node to house a possibly complicated data transformation function to consolidate the source data becomes the key problem in minimizing the total energy consumption in transmission for the data query. This dissertation investigates the mathematical model of the energy consumption in transmission for data query and the fusion node selection strategy, and proposes an Energy Efficient fusion node Selection Strategy termed EESS without knowing the global information of the network topology. Simulation results show that EESS performs better than other approaches in low-density networks and for queries requesting from data sources at a distance, thus extending the operational lifetime of the sensor network.For the circumstances that multiple sink nodes query the same source node for the same copy of data, the "one-to-many" multicast communication method can be adopted both in the routing layer and the data link layer to reduce the amount of data transmitted. In multicast communication, the same copy of data can be sent to multiple receivers at the same time by using the wireless channel, so as to avoid the same data being transmitted for several times. In the routing layer, this dissertation proves that the minimal energy multicast routing tree problem is NP-completed and proposes a distributed near-optimal Minimum Energy Routing Protocol termed MERP without knowing the global information of the network topology. In the data link layer, this dissertation extends the multicast operation of the S-MAC protocol and proposes a reliable multicast MAC protocol called mMAC by introducing and using some new control frames. Simulation results show that, the energy consumption in transmission can be saved by adopting the MERP protocol and can be further reduced by combining the MERP protocol together with the mMAC protocol.When multiple sink nodes inject different queries for the data from the same source node with different accuracies, how to disseminate the data streams to sink nodes at different query frequencies becomes a new problem. Based on this kind of application, this dissertation proposes a Multi-Rate data Query Model in WSNs and presents an Energy-efficient query processing Strategy called E-Strategy. In E-Strategy, the data stream with highest query frequency is broadcast by using the wireless channel from the source node to multiple sink nodes along the routing path. Taking into account the data correlation, the data streams at lower frequencies can be reconstructed from the consolidated stream at the highest frequency. In this manner, the energy consumption for transmitting the data streams at lower frequencies can be saved by taken tolerable errors. Both the accuracy and energy efficiency are evaluated by simulation. Simulation results indicate that the performance gain is significant and the average relative error is no greater than 5%. Besides applications with real variable source data, E-Strategy can also be extended to multimedia applications with audio, image and video source data.