Research On Wireless Sensor Network And Its Applications

Advances in Micro Electro-Mechanical System (MEMS) technologies, embedded computing technologies and wireless communication technologies have enabled the development of relatively inexpensive and low-power-consumption micro sensors with the capability of sensing, computing and communicating. Composed of a large number of these sensor nodes, a Wireless Sensor Network (WSN) can be used for detecting, collecting and analysing the information of complex environments in real time. It has a wide range of applications in military communication, environment monitoring, traffic control, personal health monitoring, home networking, etc.In this dissertation, the following topics related with WSN are investigated and some methods are proposed to improve performance of the existing methods.(a) We investigate the sensor clustering strategies of a WSN. Since a sensor usually operates on battery, the network must be power-efficient. Clustering has proven to be an effective technique to organize a large number of sensors into a connected hierarchy to reduce the energy dissipation and as well improve the efficiency of a WSN. Clustering algorithms such as LEACH, EECH, HEED, SEP DEEC and TEEN are proposed in recent literatures, which improve the performance of a WSN in various aspects such as power efficiency, heterogeneity, cluster head balance, cover model, mobility, etc.In this dissertation, we first studied the design rationale of sensor clustering approaches. The features, node capability requirements, and proper application backgrounds of different clustering approaches are discussed in detail. A novel clustering approach, which is referred to as Pre-emptive Multi-Hop Clustering Algorithm (PMHC), is then proposed. The novel features of the proposed approach include: (1) the concept of "reference energy" is proposed to measure the residual energy of a node. Cluster heads can then be determined based on the ratio of residual energy against the reference energy. This method ensures that the cluster heads are uniformly distributed, and therefore significantly reduces the iteration time. The reference energy can be computed from the neighbouring nodes in an adaptive and distributed way; (2) it establishes a multiple-hop cluster with a minimum inner-cluster communication cost. Simulations on OMNet++ rigorously validate the performance of the proposed algorithm. The experimental results show that the proposed PMHC isable to achieve a longer network lifetime and higher effective cluster-base-station data rate than the classic clustering algorithms such as LEACH and HEED.(b) Though a WSN is typically composed of a large number of homogeneous, tiny, resource-constrained, mostly immobile sensors, sensor nodes may still have varying computing, communication capabilities and different power supplies and in some networks, the sensor nodes are movable. Some research is found in literatures on such networks, in which a few resource-rich and fixed sensors act as the backbone nodes and other resource-sparse or mobile sensors act as the leaf nodes. In this dissertation, we investigated another scenario that follows. Suppose a network contains both fixed nodes and mobile nodes. The fixed nodes are usually required to operate with a low power supply since they are difficult to recharge. On the contrary, the mobile nodes are allowed to operate with a higher power since they can be easily recharged. Thus, the fixed nodes should be power-efficient. In this dissertation, a Beacon-Based Heterogeneous System (BBHS) is proposed to reduce the power consumption of the fixed nodes by taking advantage of the energy of mobile nodes. Furthermore, the operation of mobile nodes is also optimized with the help of beacons (the fixed nodes are taken as the beacons in BBHS ). This system model suits well the mine security network, visitor monitoring and guiding system for expo, and health monitoring system for athletes.(c) Most conventional Intelligent Transportation Systems (ITS) rely on the wired sensors to monitor the traffic. This wired system on one hand increases the cost of construction and maintenance. Furthermore, they can only monitor the vehicles in limited areas. Because of the advantages of WSN such as the low power consumption, wireless distribution, all-road coverage etc., the application of WSN in ITS is expected to overcome the above-mentioned difficulties. In this dissertation, a road model equipped with WSN is proposed. Based on the model, two optimization algorithms are designed to improve transportation efficiency. The first Minimal Waiting Time (MWT) algorithm attempt to optimize the signal light control and reduce the vehicles' waiting time on road intersections. The second Minimal Travel Time (MTT) algorithm attempts to reduce the whole travel time by selecting the optimal travel path dynamically, responding to the real-time traffic situation.With the introduction of WSN, the performance of the ITS can be improved further, since it enables the bidirectional communication between the roads and the vehicles. This provides a possibility for diverse personal services in future ITS.At last, this dissertation prototype of wireless sensor node. The guideline of selecting the chip, the module architecture of a node, schematics of some key module is described in this part. This work laid a foundation for further researches and applications.