Research On Energy-Efficient Algorithm For Wireless Sensor Networks

With the development of wireless communication, micro-processing and sensing techniques, it is available to deploy wireless sensor networks (WSNs) composed of a large number of sensor nodes. Both the energy consumption and storage consumption between nodes are unbalanced in WSNs. In traditional multi-hop WSNs, all nodes cooperate to transmit data to the external Sink. Sensor nodes closer to the Sink tend to forward a large number of data, which will deplete their energy faster. In data-centric storage (DCS) schemes, data are stored by name at nodes within the network. Queries for a particular name are sent to the node storing those named data. These nodes will receive a great number of data and process query requests, which will deplete their energy and storage faster.Restricted by the bulk and cost, both the node's energy and storage are limited. Sensor nodes are usually powered by battery, and the energy supply is extremely limited and hard to recharge. This becomes a bottleneck problem of wireless sensor networks. On the other hand, data lost may be caused by the limitation of storage, which lowers the data quality. It is very important to efficiently utilize the energy and storage, and prolong the network lifetime of wireless sensor networks.Based on a systematic summarization on relevant works of wireless sensor networks, this dissertation focuses on related technologies on energy hole problem in multi-hop wireless sensor networks and data-centric storage. The main contributions are as follows:1. To overcome the limitation imposed by fixed layer, this dissertation proposes MOL (Minimum Overlapping layers) scheme, a dynamic layer boundary model, to balance the energy consumption between cluster heads. Using MOL scheme, nodes are partitioned into layers according to their distance to the Sink, and a layer may be partitioned into clusters. When the residual energy of a cluster head falls below the threshold, a new round of network formation will be triggered. Some nodes may have their layer ID changed after network reformation. Consequently, throughout the network lifetime, the boundary between neighbor layers is moving towards the Sink as needed and the cluster size is dynamically changed. The MOL scheme overcomes the limitation caused by static network topology control and can adapt to any randomly deployed network as long as the initial topology is connected. The MOL scheme inherently helps balance the energy consumption among cluster heads and improve the network lifetime.2. This dissertation studies the energy hole problem in many-to-one multi-hop WSNs. First, the traffic load analysis in the continuous space of the network is performed, which reflects the gradual change of the dynamic layer boundary. Next, the impact of idle listening to the average load is studied. Based on the analysis, a novel load-similar node distribution strategy combined with the MOL scheme is proposed to alleviate the energy hole problem in multi-hop layered WSNs. Sensor nodes are deployed in the network area according to the load distribution, i.e., more nodes will be deployed in the area where the average load is higher, then the load among different areas in the sensor network tends to be balanced. Compared with an existing nonuniform node distribution and uniform node distribution strategies, the proposed load-similar node distribution strategy achieves good energy balance among different layers in the network and prolongs network lifetime. Note that although this dissertation demonstrates the load-similar node distribution strategy is best suitable for dynamic layered WSNs, the analysis model and the proposed load-similar distribution strategy actually can be well applied to other multi-hop WSN structures.3. To address the hotspot problem of data-centric storage in wireless sensor networks, this dissertation proposes an energy-efficient hotspot-aware data storage scheme (SASS). When a sensor node detects an event, the storage location of the event is calculated based on the event type, and the event is routed to the closest node (called home node) of the storage location. When the home node depletes the energy or storage, new storage node is selected dynamically according to the neighbors'residual energy and storage to avoid data lost caused by hotspot problem. The routing process is improved to reduce communication cost caused by perimeter forwarding and reduce energy consumption of nodes in hotspot area. SASS can effectively alleviate the hotspot problem; reduce data lost and energy consumption.