Researches On Data Gathering And Survival Algorithms In Wireless Sensor Networks

Wireless sensor networks (WSNs), which are composed of a large number of low-cost, low-power, multifunctional sensor nodes, possess advantages such as fast deployment, adaptation to various challenged environments, accurate and credible data sensing and collecting in covered network regions. So WSNs have wide prospects for military, environmental, health, transportation, and commercial applications. On one side, node resources in WSNs, such as processing ability, buffer size, bandwidth and energy, are strictly limited due to cost control, volume control, and other reasons. In order to reduce resource consumption, sense nodes often adopts special scheduling models of periodic working/sleeping and short radio range. Thus sensor nodes cooperate each another to forward collected data to sink nodes through multi-hop transmission. It generates the features of opportunistic connectivity and multi-hop transmission in WSNs. On the other hand, WSNs are often deployed in harsh, even extreme environments, which may make sensor nodes smashed or invalidated by the destructive factors of environment. Moreover, sensor nodes have limited energy, and are very hard to be recharged. A sensor node will die once its energy is exhausted. Thus, the data gathering problem caused by opportunistic connectivity and multi-hop transmission, and the data survival problem caused by node corruption and invalidation, are crucial in the large-scale implemention of WSNs, and also become the key points and the chief difficulties in the research of WSNs.This dissertation involves regular and comprehensive study and analysis on wireless sensor networks consist of static nodes and delay tolerant mobile sensor networks (DTMSNs) proposed in recent years, and, on this basis, carries out in-depth research in data gathering and data survival problems in WSNs. The original achievements and contributions of the dissertation are highlighted as follows:1. To save node energy and prolong network life while gathering data from nodes to sink nodes, an Energy-efficient DAta Gathering (EDAG) algorithm for DTMSN is proposed. Based on community mobility models, EDAG takes two techniques as follows. First, the proposed protocol calculates the delivery probability of each node based on both the frequency, which the node meets with sinks, and its mobility trend. Second, EDAG extends the eyeshot of each node when it looks for next hop through finding and using the connected paths formed dynamically by mobile sensor nodes. So EDAG achieves high data delivery ratio closed to current multiple copy protocols but by single copy transmission which has notable energy saving advantages. Simulation results have shown that the proposed EDAG achieves the comparable message delivery ratio with the much lower transmission overhead than several main data delivering approaches for DTMSNs. What's more, EDAG can efficiently save node energy and remarkably prolong network life, which make it well fit for the energy-limited characteristic of DTMSNs well.2. A new data gathering scheme RADG (an Replicas Adaptive Data Gathering Scheme) for DTMSNs is proposed. Due to the inherent feature of intermitted connectivity and varying topology, DTMSNs usually gather data by the way of probabilistic forwarding. It is reasonable that DTMSNs routing employs multi-copy schemes to improve the message delivery ratio and reduce the delay, considering that probabilistic forwarding can not ensure good performance. However, the approach of injecting a large amount of message copies into the network will drain the limited network resource of DTMSNs including bandwidth, battery supply and storage space. So a proper routing method need to trade off between the number of copies of messages and the network performance. The proposed RADG economizes network resource using a self-adapting algorithm to cut down redundant copies of messages, and achieves a good network performance by leveraging the delivery probabilities of the mobile sensors as main routing metric. Simulation results have shown that RADG achieves the higher message delivery ratio with the lower transmission overhead and data delivery delay than other DTMSNs data delivering schemes.3. A fast and efficient data survival algorithm called FEDS for WSNs is proposed. For devastating events usually only affects a limited area, there exists a big difference in the security status of sensor nodes. Moreover, there is a gap from the time when the disaster hits and the time when the time a significant portion of nodes are physically destroyed. Thus, for achieving high data survival ratio, a feasible scheme is to transfer as much data as possible from nodes in dangerous status to those in security status, preferably within as a short period as possible. According to the information collected of sensor nodes in security status, FEDS can calculate the optimal solution of data transfer according to the“Transportation Problem”in integral linear programming theory. Simulation shows that FEDS achieves high data survival ratio through fast and efficient data transfer.4. A Virtual Gravity based Data Survival (VGDS) algorithm is proposed for unattended wireless sensor networks, in which there are not continuously available sink nodes. VGDS is a distributed algorithm, i.e., data transfer is carried out by sensor nodes in a peer-to-peer cooperative way. To improve data survival ratio and reduce evacuation time, VGDS adopts virtual gravity method to calculate the near optimum solution of data transfer. The simulation results shows that VGDS achieves high data survival ratio at acceptable time price, and effectively guarantees high data survival ratio in the networks in serious environments.