Research On Periodic Data Gathering In Wireless Sensor Networks

The emergence of wireless sensor networks is making peoples' lives more convenient, and helps them more efficiently connect with the physical world. Wireless sensor networks are usually deployed in tougher outdoor environments where there is the absence of care. The capabilities of node processing and communication are limited, but people need them to provide more effective and longer service. Due to strict energy constraints, sensor network's energy resources should be carefully managed to extend the lifetime of the network. In many applications of wireless sensor networks, the sensed data must be periodically gathered and transmitted to a comparatively remote base station. This process is called data gathering which involves networking, routing, data fusion and other key technologies, and is a comprehensive issue.In this dissertation, several key issues in periodic data gathering are studied. These issues are energy equalizing, energy efficiency and delay balance, data-aware clustering. The resolved problems are "bad-links" problem, "hot spots" problem, latency optimization problem, the rational use of data correlation. The proposed protocols are also achieved and the performances are verified. The main work and contributions of this dissertation include:1. Proposes an energy equalizing data gathering schemeThe causes of energy imbalance, such as the random distribution of nodes, the irregular of the monitoring information in time and space are analyzed and summarized. A low delay energy equalizing routing scheme LDEERS is proposed for data gathering. It consists of rapid establishment phase, topology adjustment phase and data collection phase. LDEERS introduces a dynamic route selecting scheme based on weight-value and height-value to form a tree and a mechanism similar to token passing to elect the root of the tree in each round. In each round, the current root node collects and integrates all the information sent from nodes and transmits to the base station, and then selects the next root node. Underpinned by these schemes, LDEERS can simply and fast organize all the nodes with low overhead and is robust enough to the topology changes. Compared with PEGASIS (Power-Efficient Gathering in Sensor Information Systems) and HEED(Hybrid, Energy-Efficient, Distributed clustering approach), simulation results show that LDEERS achieves its goals of consuming less energy, equalizing the energy consumption of all the nodes, as well as extending the network lifetime perfectly.2. Proposes an energy efficiency and delay balancing data gathering schemeTo consider the data gathering delay while seeking high energy efficiency in wireless sensor networks, energy efficiency and delay balancing data gathering scheme EEDBDG is proposed. EEDBDG organizes the topology based on a novel dynamic tree and avoids the "hot spots" problem. Nodes select routing path dynamically and take turns to act as the root of the tree. The root node gathers the data and communicates with the base station directly. In EEDBDG, three data gathering strategies are proposed for different request of delay and energy efficiency. These are optimum energy efficiency algorithm EEDBDG-E(Energy), optimum delay algorithm EEDBDG-D(Delay) and a compromise algorithm EEDBDG-M(Median). In the case that the sensor nodes communication radius is restricted, simulation results show that EEDBDG equalizes the energy consumption of all the nodes, prolongs the network lifetime and acquires an outstanding performance in energy-saving and time-thrift. Compared with GSEN(Group-based SEsor Network), in the best case, the network lifetime of EEDBDG-E is prolonged by 72% and the data gathering delay using EEDBDG-D is decreased by 74%. The security issues of dynamic tree-based data collection are also discussed, and a security scheme is proposed.3. Proposes a data-aware clustering for data gatheringThe data fusing ability of a cluster head is affected by the correlation of the sensed data from different nodes in wireless sensor networks, which are data centered. The influence of regional data correlation on the energy efficiency of a clustering protocol is analyzed, and it is pointed out that the assumptions of existing protocols for data correlation and data fusion capability of cluster head are one-sided and too ideal. A data-aware clustering for data gathering DACDG is proposed. DACDG uses the residual energy and the correction of sensed data from the neighbors to construct clusters, and adopts iterative computing to elect a cluster head. When compared with existing protocols, our simulation results show that DACDG achieves its goals of consuming less energy, equalizing the energy consumption of all the nodes, and extremely extending the network lifetime.4. Designs an experimental system to verify the performances of data gathering protocolsAn experimental system of data collection is designed. An embedded protocol stack was designed and used in our system, including the Radio Frequency driver layer, MAC(Medium Access Control) layer, network layer and transport layer. The protocol stack adapts to most of the current wireless sensor network node platform, which can quickly construct the network systematically, and realize data communication. The data gathering scheme EEDBDG is achieved in the protocol stack. An effective stable data communication networking approach called MIN (Multiple Interactive Networking) is designed to establish a quasi-stable link and to form the network. MIN transmits data using the most reliable node, improves the reliability of data transmissions, and prefects the establishment of data aggregation tree. Finally, the test results verify the performance of the protocol.