| Wireless sensor networks (WSNs) are the multi-hop and self-organizing networksystems. A wireless sensor network is composed of dozens or even thousands of miniaturewireless sensor nodes that are deployed in monitoring region and complex environment.WSNs use different short-range wireless communications technologies for differentapplications. With wireless communication technologies developing rapidly, people couldgain information by WSNs at anytime, anywhere, and any environment. Therefore, this isthe foundation of the Internet of Things (IoT). As an important part of the IoT underlyingnetwork, concerning each object of the IoT, ability to detect changes in the physical statusof things is also essential for recording changes in environment. In this regard, WSNs playa crucial role in bridging the gap between the physical and virtual world, and enablingthings to respond to changes in their physical environment. However, in a distributedsystem, such as WSNs and IoT, the performance of network models decides network lifeand applications range at first. Secondly, since crystal frequency of different nodesgradually deviate and a distributed system generally requires better time synchronizationaccuracy, it is significant to work out effective time synchronization mechanism under aideal model. At last, how to merge WSNs techniques into Internet techniques based on IPprotocol is also a key problem to be solved for IoT.Efficient and low-power network model and the time synchronization mechanism hasbecome an emerging issue for WSNs research areas. Moreover, mutual integration betweenWSNs and Internet based on IP will be upcoming research points of future IoT development. By in-depth study on WSNs models, time synchronization mechanism andtechnical characteristic of IoT, this paper analyzes performance indicators of WSNsmodeling, summarizes advantages and disadvantages of different types of timesynchronization mechanism as well as the main factors affecting the synchronizationaccuracy, and concludes the developing process and features of the IoT. In WSNs, exceptsink node and few nodes, all others must exchange time information by the effective timesynchronization mechanism to keep time synchronization with each other. However,current time synchronization mechanism for WSNs has some problems to be solved, suchas high algorithm complexity, long time synchronization convergence time and sensitiveerror data. The core idea of mutual integration between underlying WSNs and Internet is toform a complete IP network; IP-optimized network is an important technologicalbreakthrough in terms of heterogeneous networks integration. For IoT, based on IP,addressing strategy could provide services for a variety of upper applications, andsimultaneously allow heterogeneous network based on IP access IoT smoothly.Aiming at the existing problem in current WSNs and IoT, some creative ideas andsolutions are proposed, the innovation and contribution of the paper lie in the followingfive aspects:1. Based on analyzing IoT two basic concepts, architecture of IoT are proposed, whichinclude Underlayer Network Distribution, Convergence Gateway Access, Inter-connectednetwork Integration and Terminal User Application. In the architecture, a protocol structureof IoT is designed, which consists of Network Protocol Layers, Network Control Platformand Application Terminal Platform, and the key technologies for IoT have been discussedconcerning hardware and software. Six development conceptions of future IoT have beenpresented based on summarizing the existing problems of IoT in standards, technologies,security and application. A model of production quality management for auto parts basedon IoT has been explored and established.2. Modeling for wireless sensor networks is very challenging because the modelingneeds to adapt network dynamics and find out multiple optimizing paths fromthe microscopic point of view. In this paper, we propose a cellular automata model thatfocuses on dynamic network topology, multipath data transmission mechanism and energyoverhead. Each node in a network is represented by a cell, any links between two cellsconstitute the cellular space. A wireless sensor network is modeled by related cell statesand cell evolution rule. A cell transmission model is derived for multipath transmission ofinformation. All these aim to ensure network reliability with the minimum resourcerequirements. Presented analytical work is proved validly by simulations. 3. The traditional Reference Broadcast Synchronization (RBS) algorithm has aproblem that network overhead is very large as the network nodes increase in Wirelesssensor networks. An Energy-efficient RBS (ERBS) scheme is presented to work out thenetwork overhead issue. Firstly, every receiving node, which is required to receive a coupleof reference messages, computes mean phase offset to its nonadjacent receiving nodes, andestimates phase offset by maximum posteriori estimation; Secondly, the algorithm usesleast-squares linear regression to fit clock skew periodically. The analysis on simulationresult indicates that ERBS algorithm improves synchronization precision and reducesenergy usage over RBS.4. The current time synchronization algorithms for WSNs have some defects, whichinclude sensitive error data, huge energy consumption and long synchronization convergencetime. The Flooding Broadcast Time Synchronization algorithm based on voronoi diagram(FBTS) is presented to work out related issues. The basic synchronization idea of thealgorithm is to record a broadcast message with time-stamp between sender and receiver,and the data are clustered around by K-means method, and then linear regression is used tocompensate clock drift after removing the data deviated from the normal error range.Furthermore, to meet large-scale networks requirement, multi-hop network timesynchronization mechanism is constructed by the virtual synchronization root node basedon voronoi synchronization model. The analysis on simulation result indicates that FBTSalgorithm improves synchronization precision, reduces energy consumption andsynchronization convergence time compared with other algorithms.5. Addressing scheme for IoT is proposed based on6LoWPAN, for the real-timecommunication between Internet and underlying heterogeneous networks based on IEEE802.15.4. Addressing scheme includes IPv6address autoconfiguration and headcompression. Hierarchical address autoconfiguration firstly allows nodes use link localaddress, which is derived by16-bit short address, for transmitting data packets withinWPAN, meanwhile, this link local address needs to ensure the uniqueness of16-bit shortaddress by performing duplicate address detection based on clustering. Secondly, all thesink nodes in underlying networks form global address by obtaining global address prefix,which combine with interface identifier, and achieve data exchange between Internet andthe WSNs. Simultaneity, IOT IPv6Header Compression (IIPHC) is proposed byembedding G/L bit that decides which types of header compression scheme is adopted inhead compression encoding. For the link local address, simple IIPHC1program is used.For the global address, IIPHC2that is relatively complex but effective program is used.The simulation result indicates that addressing scheme has an improvement in networks overhead, latency, throughput and energy consumption.In summary, the dissertation solves the key problems existing in the WSNsTechnology for IoT from multiple perspectives, by researching on the IoT technology andapplications, WSNs network models, WSNs time synchronization mechanism and IoTaddressing scheme, and play a role in promoting IoT application and development based onWSNs. |
PDF Full Text Request