Research On Energy-Efficient Data Aggregation Protocols In Wireless Sensor Networks

Recent advances in wireless communications, electronics and micro-sensor technologies have enabled the development of small, low-cost sensors with sensing, computation and wireless communication capabilities. Wireless Sensor Networks (WSN) are composed of a large number of sensor nodes and one or more sink nodes, and sensor nodes can form a multi-hop communication network automatically. Hence, WSN can be deployed conveniently and quickly, and are robust to failure of nodes. How to prolong the lifetime of wireless sensor networks is an important issue in designing environmental monitoring applications.In order to tackle this problem, we propose an energy-efficient polynomial regression model for data gathering in this paper. Each sensor node in the network fits a regression function with its sensed data in most recent rounds, and sends coefficients of the regression function and some related parameters to the sink node instead of sending the sensed data. Based on the model, we improve it by selecting feature data to further reduce the communication cost. Namely, sensor node selects a set of feature points of regression polynomial, and sends the data of the points to the sink node. The sink node generates the regression polynomial according to the receiving data and approximates the sensing data. Theoretical analysis and simulation studies show that the proposed algorithm can greatly reduce the number of data transmissions among the sensor nodes while keeping high accuracy of sensory data, leading to significant energy saving on the sensor nodes.In this paper, we also propose a novel energy-efficient data gathering algorithm, which is based on dual prediction. Both sensor nodes and sink node use the same prediction algorithm based on the same historical data. Each sensor node predicts the values of the data to be sensed according to its recorded historical information. When getting the sensed data, the node compares it with the predicted data. Only when the difference between the predicted and the sensed data values exceeds a pre-defined threshold, the sensor node sends the newly sensed data to the sink node. Theoretical analysis and simulation studies show that EDP can greatly reduce the amount of data transmitted in sensor nodes, leading to significant energy saving on the sensor nodes and extension of the lifetime of the entire network.Many data gathering schemes have been proposed in the literature. To save energy of sensor nodes, those schemes utilize various routing techniques to select energy-saving paths for data packets, and data aggregation techniques to reduce spatial-temporal correlation of the sensed data. In fact, the energy can be saved further. This paper presents two data aggregation schemes to reduce the temporal redundancy of sensory data at the source. The proposed protocols in this paper provide several schemes which aggregate the temporal related sensory data, which will be valuable for other researchers in designing data aggregation protocols in wireless sensor networks.