Primary Initialization For Wireless Sensor Networks

A wireless sensor network(WSN), consisting of a large number of sensor nodes,is a multi-hop self-organizing network system. As most of the wireless sensor networks were deployed manually at the beginning, the location and the neighboring relationships can be set accurately. Thus it is unnecessary to think about the self organizing problems for the newly deployed networks. With expanding of the scales and extending the application fields of WSNs, more and more applications have to be deployed by scattering the sensor nodes by the helicopters or canons arbitrarily. In the unstructured phase decided by such deployment method, all the sensor nodes know neither which nodes are their neighbors nor how many neighbors they have. Therefore, there must be a primary initialization process to build up the initial network structure before the network can really work. With the development of WSNs, the primary initialization becomes an urgent research project, which is the important process for the implement of WSNs.The co-channel interference, which is induced by the characteristic of no structure at the primary initialization phase, decreases seriously the reliability of data transmissions between nodes. As a result, not only the existing MAC (Medium Access Control) protocols are not available to the primary initialization phase, but also the traditional initializing algorithms based on the networks with the complete structures are not applicable. To solve the primary initialization problem should combine the following two aspects: the data transmissions problem at the primary initialization phase of WSNs; the problem of building the topology for the newly deployed networks. Based on that, the dissertation focuses mainly on the data transmission between nodes and the core algorithms for the primary initialization, as well as the functional evaluation indexes and the applicable techniques for the primary initialization algorithms.The contributions of the dissertation include:(1) Analyze the data transmission rules during the primary initialization phase. Demonstrate critically the co-channel interference of high density sensor nodes with equal sending probability under SNIR (Signal to Noise and Interference Ratio) co-channel model. Two theorems: Single-Node to Single-Node Transmission Failure Theorem and Single-Node to Multiple-Nodes Successful Sending Theorem are proposed and proved critically, which provide the theoretical basis for the communication of the primary initializing process.(2) Prove theoretically the core algorithm of the primary initialization for the sensor networks. Propose the clustering strategy which is communication-related and competence-based; bring forward the coloring-based time-slots assignment principle; design the complete TDMA(Time Division Multiple Access)-based single-channel primary initialization algorithm; prove theoretically the correctness and completeness of the algorithm; validate the algorithm by the simulations.(3) Propose the applicable primary initialization algorithm and substantiate the algorithm by the large number of simulations. Put forward the three-channel applicable technique aiming to further decrease the co-channel interference; propose the adaptive technique which adjust the sending probability by the local environmental parameters; design the clustering-based multi-channel primary initialization algorithm; build up the evaluation indexes for the algorithm. By a lot of simulation experiments with the different network scales, the different network densities and the different sending probabilities, it is fully proved that all evaluation indexes can satisfy the requirements of the applications. Besides, the effective referent parameters of the algorithm are obtained from the simulation results. All the work establishes the valuable experimental basis for the primary initialization algorithms.The achievements of the dissertation can be used to not only the theoretical analysis for the newly deployed wireless sensor networks but also the applications of primary initialization engineering for WSNs.