| The powerful data processing ability of sensor networks over wireless communication environment has attracted much focus from both academic and industrial fields. A key issue in this complicated and volatile sensor environment is to timely and effectively obtain the monitoring information, which has increasingly become an active research topic in recent years. Recent advance in data fusion technology has enabled the necessity of interoperability between the vertical data system, seamless communication, as well as data sharing. In this dissertation, we design the architecture of wireless sensor network based information exchange platform, which aims to achieve the effectively sharing of heterogeneous data, together with interaction of the general information. In this paper, we focus on the data gathering strategies concentrating on the information exchange platform. Note that the main work covers the following aspects:In view of the native data characteristics of multi-streaming and isomerism in wireless sensor networks, it is necessary to build a common, service-oriented data integration platform, the objective of which is to resolve the problem of how to access the data, which application is interested in the data, and how to use the data. In this way, the platform abstracts the information provider from different users, which made the user application programs not disturbed by the interest distribution in sensor networks. With considering the characteristics of isolation of sensed resources, this information exchange platform can realize the perception data flow interaction over the application layer.To save node energy and prolong network lifetime in the course of gathering data from source nodes to the sink, an energy-efficient data gathering algorithm is proposed. For the purpose that multicast tree based routing well balances the data transmission delay and network bandwidth consumption, it takes advantage of much longer shared paths to cut down the cost for the constructed spanning tree. In addition, our proposed method brings the concept of energy level into this fast low-cost shortest path tree algorithm, which takes full consideration of the node energy level while chooses the optimal router. The result of simulation output and analysis shows that the proposed method could well lengthen the network survival. Focusing on the gradient-based routing, a novel maximizing energy utilization data gathering routing protocol (MEURP) was developed. In order to relieve the issues of network congestion and channel contention brought by the flooding mechanism, a back-off waiting scheme is also proposed, which lead to the flooding packets walking down in the increase direction of gradient field. Together, MEURP also adopts a routing table-driven data transfer mode to keep the sensed interests forwarding along the shortest gradient path. Simulations make complete comparisons between proposed back-off waiting scheme and traditional flooding mechanism over the performance of the route setup time and network load, it is clear that our proposed MEURP achieves a good balance between network throughput and system survival time.Responding to the massive data flow in wireless sensor networks, a Kalman prediction based data collection algorithm is proposed. Based on the correlation existing in both spatial extension and temporal duration, we first formulate the data stream model, and then present a Kalman filter based prediction operation. Effective error analysis model guarantees that this proposed algorithm can well hold the credible prediction accuracy. With resorting to the probability graph theory, we formulate this wireless transmission network as a factor graph. By means of the novel belief propagation algorithms, we can obtain the reliability of each transmission link, and then a greedy algorithm based data collection method is established. It is clear that this probabilistic reasoning based data collection algorithm shares high performance of network throughput.
|
PDF Full Text Request